id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2501.12657
|
Sai Duan
|
Yu Zhu, Renjie Xue, Hao Ren, Yicheng Chen, Wenjie Yan, Bingzheng Wu, Sai Duan, Haiming Zhang, Lifeng Chi, Xin Xu
|
Reconstructing Pristine Molecular Orbitals from Scanning Tunneling Microscopy Images via Artificial Intelligence Approaches
|
19 pages, 4 figures, 4 extended data figures
|
JACS Au, 2025, 5 (7), pp 3163-3170
|
10.1021/jacsau.5c00310
| null |
physics.chem-ph physics.app-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Molecular orbital (MO) is one of the most fundamental concepts for molecules, relating to all branches of chemistry, while scanning tunneling microscopy (STM) has been widely recognized for its potential to measure the spatial distribution of MOs. However, the precise characterization of MO with high resolution in real space is a long-standing challenge owing to the inevitable interference of high-angular-momentum contributions from functionalized tips in STM. Here, leveraging advances in artificial intelligence for image recognition, we establish a physics-driven deep-learning network, named STM-Net, to reconstruct MOs from high-resolution STM images with a functionalized tip, taking advantage of the separable characteristics of different angular momentum contributions. We demonstrate that STM-Net can be directly applied to a variety of experimental observations, successfully reconstructing pristine MO features for molecules under diverse conditions. Moreover, STM-Net can adapt to various states of the functionalized tip and the substrate, illustrating the broad applicability of our physics-driven framework. These results pave the way for accurate characterization of MO with high resolution, potentially leading to new insights and applications for this fundamental concept in chemistry.
|
[
{
"version": "v1",
"created": "Wed, 22 Jan 2025 05:39:46 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhu",
"Yu",
""
],
[
"Xue",
"Renjie",
""
],
[
"Ren",
"Hao",
""
],
[
"Chen",
"Yicheng",
""
],
[
"Yan",
"Wenjie",
""
],
[
"Wu",
"Bingzheng",
""
],
[
"Duan",
"Sai",
""
],
[
"Zhang",
"Haiming",
""
],
[
"Chi",
"Lifeng",
""
],
[
"Xu",
"Xin",
""
]
] |
2501.13549
|
Oleksandr Tsyplyatyev
|
Henok Weldeyesus, Pedro M.T. Vianez, Omid Sharifi Sedeh, Wooi Kiat Tan, Yiqing Jin, Mar\'ia Moreno, Christian P. Scheller, Jonathan P. Griffiths, Ian Farrer, David A. Ritchie, Dominik M. Zumb\"uhl, Christopher J.B. Ford, Oleksandr Tsyplyatyev
|
Dominant end-tunneling effect in two distinct Luttinger liquids coexisting in one quantum wire
|
30 pages, 9 figures
|
Nature Communications 16, 6997 (2025)
|
10.1038/s41467-025-62325-5
| null |
cond-mat.str-el cond-mat.mes-hall quant-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Luttinger liquids occupy a special place in physics as the most understood case of essentially quantum many-body systems. The experimental mission of measuring its main prediction, power laws in observable quantities, has already produced a body of exponents in different semiconductor and metallic structures. Here, we combine tunneling spectroscopy with density-dependent transport measurements in the same quantum wires over more than two orders of magnitude in temperature to very low electron temperatures down to $\sim$40 mK. This reveals that, when the second 1D subband becomes populated, the temperature dependence splits into two ranges with different exponents in the power-law dependence of the conductance, both dominated by the finite-size effect of the end-tunneling process. This result demonstrates the importance of measuring the Luttinger parameters as well as the number of modes independently through spectroscopy in addition to the transport exponent in the characterization of Luttinger liquids. This opens a new pathway to unambiguous interpretation of the exponents observed in quantum wires.
|
[
{
"version": "v1",
"created": "Thu, 23 Jan 2025 10:50:25 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 11:53:43 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Weldeyesus",
"Henok",
""
],
[
"Vianez",
"Pedro M. T.",
""
],
[
"Sedeh",
"Omid Sharifi",
""
],
[
"Tan",
"Wooi Kiat",
""
],
[
"Jin",
"Yiqing",
""
],
[
"Moreno",
"María",
""
],
[
"Scheller",
"Christian P.",
""
],
[
"Griffiths",
"Jonathan P.",
""
],
[
"Farrer",
"Ian",
""
],
[
"Ritchie",
"David A.",
""
],
[
"Zumbühl",
"Dominik M.",
""
],
[
"Ford",
"Christopher J. B.",
""
],
[
"Tsyplyatyev",
"Oleksandr",
""
]
] |
2501.14690
|
Athanasios Koutsostathis
|
Athanasios Koutsostathis (1), Claude Krantz (2), Elena-Oana Hanu (2), Frank Herfurth (2), Regina He{\ss} (2), Wolfgang Kaufmann (2), Michael Lestinsky (2), Mirko Dieter Looshorn (3,4), Esther Babette Menz (2), Andreas Reiter (2), Jon Ro{\ss}bach (2), Stefan Schippers (3,4), Phe Man Suherman (2) ((1) Department of Physics, National Technical University of Athens, Greece, (2) GSI Helmholtzzentrum f\"ur Schwerionenforschung, Darmstadt, Germany, (3) I. Physikalisches Institut, Universit\"at Gie{\ss}en, Giessen, Germany, (4) Helmholtz Forschungsakademie Hessen f\"ur FAIR (HFHF), Campus Gie{\ss}en, GSI Helmholtzzentrum f\"ur Schwerionenforschung, Giessen, Germany)
|
Upgrades and maintenance of the CRYRING@ESR electron cooler for improved internal electron target operation
|
This work was presented with a poster at the 32nd Annual Symposium of the Hellenic Nuclear Physics Society HNPS2024, in Thessaloniki. This article has been submitted to the 31st volume of the HNPS Advances in Nuclear Physics proceedings journal
|
HPNS Adv. Nucl. Phys. 31, 86 (2025)
|
10.12681/hnpsanp.7988
| null |
physics.acc-ph nucl-ex physics.atom-ph physics.ins-det
|
http://creativecommons.org/licenses/by/4.0/
|
The electron cooler of the CRYRING@ESR storage ring at the GSI-FAIR accelerator complex is a unique instrument, which not only provides beam cooling of the stored ions, but also serves as a low-energy electron target for Dielectronic Recombination experiments. The minimisation of vacuum contamination and the response to rapid energy changes are key requirements of the cooler in electron target mode. Therefore, a test bench was prepared to study the outgassing behaviour of the electron gun components and a setup was constructed to evaluate the drift-electrode-modulation of the acceleration voltage of the cooler. The vacuum studies showed that the electron gun cathode was severely malfunctioning and resulted in its replacement. The drift-electrode-modulation of the acceleration voltage showed significant improvements compared to direct modulation of the terminal voltage of the cooler.
|
[
{
"version": "v1",
"created": "Fri, 24 Jan 2025 18:03:46 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Koutsostathis",
"Athanasios",
"",
"Department of Physics, National Technical University of Athens, Greece"
],
[
"Krantz",
"Claude",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Hanu",
"Elena-Oana",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Herfurth",
"Frank",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Heß",
"Regina",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Kaufmann",
"Wolfgang",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Lestinsky",
"Michael",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Looshorn",
"Mirko Dieter",
"",
"I. Physikalisches Institut, Universität Gießen, Giessen, Germany",
"Helmholtz Forschungsakademie Hessen für FAIR"
],
[
"Menz",
"Esther Babette",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Reiter",
"Andreas",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Roßbach",
"Jon",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
],
[
"Schippers",
"Stefan",
"",
"I. Physikalisches Institut, Universität Gießen, Giessen, Germany",
"Helmholtz Forschungsakademie Hessen für FAIR"
],
[
"Suherman",
"Phe Man",
"",
"GSI Helmholtzzentrum für Schwerionenforschung, Darmstadt, Germany"
]
] |
2501.15347
|
Farah Baracat
|
Farah Baracat, Luca Manneschi and Elisa Donati
|
Heterogeneous Population Encoding for Multi-joint Regression using sEMG Signals
| null | null |
10.1109/NICE65350.2025.11065840
| null |
cs.HC eess.SP
|
http://creativecommons.org/licenses/by/4.0/
|
Regression-based decoding of continuous movements is essential for human-machine interfaces (HMIs), such as prosthetic control. This study explores a feature-based approach to encoding Surface Electromyography (sEMG) signals, focusing on the role of variability in neural-inspired population encoding. By employing heterogeneous populations of Leaky Integrate-and- Fire (LIF) neurons with varying sizes and diverse parameter distributions, we investigate how population size and variability in encoding parameters, such as membrane time constants and thresholds, influence decoding performance. Using a simple linear readout, we demonstrate that variability improves robustness and generalizability compared to single-neuron encoders. These findings emphasize the importance of optimizing variability and population size for efficient and scalable regression tasks in spiking neural networks (SNNs), paving the way for robust, low-power HMI implementations.
|
[
{
"version": "v1",
"created": "Sat, 25 Jan 2025 23:15:23 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Baracat",
"Farah",
""
],
[
"Manneschi",
"Luca",
""
],
[
"Donati",
"Elisa",
""
]
] |
2501.15544
|
Hanwen Zhang
|
Hanwen Zhang, Ruichen Zhang, Wei Zhang, Dusit Niyato, Yonggang Wen, Chunyan Miao
|
Advancing Generative Artificial Intelligence and Large Language Models for Demand Side Management with Internet of Electric Vehicles
|
11 Pages
| null | null | null |
cs.LG cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Generative artificial intelligence, particularly through large language models (LLMs), is poised to transform energy optimization and demand side management (DSM) within microgrids. This paper explores the integration of LLMs into energy management, emphasizing their roles in automating the optimization of DSM strategies with Internet of electric vehicles. We investigate challenges and solutions associated with DSM and explore the new opportunities presented by leveraging LLMs. Then, we propose an innovative solution that enhances LLMs with retrieval-augmented generation for automatic problem formulation, code generation, and customizing optimization. We present a case study to demonstrate the effectiveness of our proposed solution in charging scheduling and optimization for electric vehicles, highlighting our solution's significant advancements in energy efficiency and user adaptability. This work underscores the potential of LLMs for energy optimization and fosters a new era of intelligent DSM solutions.
|
[
{
"version": "v1",
"created": "Sun, 26 Jan 2025 14:31:03 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Feb 2025 09:08:55 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 11:09:11 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 03:20:12 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhang",
"Hanwen",
""
],
[
"Zhang",
"Ruichen",
""
],
[
"Zhang",
"Wei",
""
],
[
"Niyato",
"Dusit",
""
],
[
"Wen",
"Yonggang",
""
],
[
"Miao",
"Chunyan",
""
]
] |
2501.16325
|
Declan Alex Norton
|
Declan A. Norton, Edward Ott, Andrew Pomerance, Brian Hunt, and Michelle Girvan
|
Tailored Forecasting from Short Time Series via Meta-learning
|
23 pages, 12 figures
| null | null | null |
cs.LG nlin.CD physics.comp-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Machine learning models can effectively forecast dynamical systems from time-series data, but they typically require large amounts of past data, making forecasting particularly challenging for systems with limited history. To overcome this, we introduce Meta-learning for Tailored Forecasting using Related Time Series (METAFORS), which generalizes knowledge across systems to enable forecasting in data-limited scenarios. By learning from a library of models trained on longer time series from potentially related systems, METAFORS builds and initializes a model tailored to short time-series data from the system of interest. Using a reservoir computing implementation and testing on simulated chaotic systems, we demonstrate that METAFORS can reliably predict both short-term dynamics and long-term statistics without requiring contextual labels. We see this even when test and related systems exhibit substantially different behaviors, highlighting METAFORS' strengths in data-limited scenarios.
|
[
{
"version": "v1",
"created": "Mon, 27 Jan 2025 18:58:04 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 06:57:11 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Norton",
"Declan A.",
""
],
[
"Ott",
"Edward",
""
],
[
"Pomerance",
"Andrew",
""
],
[
"Hunt",
"Brian",
""
],
[
"Girvan",
"Michelle",
""
]
] |
2501.16415
|
Pradip Kattel
|
Yicheng Tang, Pradip Kattel, J.H. Pixley, and Natan Andrei
|
Two Channel Kondo behavior in the quantum XX chain with a boundary defect
|
5 pages, 4 Figures
|
Phys.Rev.B 112, L020303, 2025
|
10.1103/t78z-pyfr
| null |
cond-mat.str-el cond-mat.stat-mech hep-th math-ph math.MP quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
We demonstrate that a boundary defect in the single spin-$\frac{1}{2}$ quantum $XX$ chain exhibits two-channel Kondo physics. Due to the presence of the defect, the edge spin fractionalizes into two Majorana fermions, out of which one decouples, and one is overscreened by the free fermion in bulk, leading to non-trivial boundary behavior characteristic of the two-channel Kondo model. When the ratio of boundary to bulk coupling exceeds a critical value of $\sqrt{2}$, a massive boundary-bound mode is exponentially localized near the impurity site for strong impurity coupling. This leads to unusual behavior in physical quantities, such as the $g$-function not being monotonic. We compute the $g-$function of the impurity from both thermodynamic and entanglement entropy calculations and show that it takes a non-integer value of $\sqrt{2}$ just as in the two-channel Kondo problem.
|
[
{
"version": "v1",
"created": "Mon, 27 Jan 2025 19:00:00 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 01:09:28 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Tang",
"Yicheng",
""
],
[
"Kattel",
"Pradip",
""
],
[
"Pixley",
"J. H.",
""
],
[
"Andrei",
"Natan",
""
]
] |
2501.16563
|
Philipp Bader
|
Philipp Bader
|
Comparing Teichm\"uller and curve graph translation lengths
| null | null | null | null |
math.GT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
A pseudo-Anosov mapping class acts on Teichm\"uller space $\mathcal{T}$ as well as on the curve graph $\mathcal{C}$ with so called north-south dynamics. We can measure a stable translation length $l_\mathcal{T}$ and $l_\mathcal{C}$ of the respective actions. Boissy and Lanneau compute the minimal Teichm\"uller translation length over all pseudo Anosovs in a fixed genus that lie in a hyperelliptic component of translation surfaces. In particular, this minimum is always greater than $\log(\sqrt{2}),$ independently of the genus. Here, we show that the minimal stable curve graph translation length over the same family of pseudo-Anosovs behaves differently: Namely, for a genus $g$ surface this minimal translation length is of order $\frac{1}{g}.$ To prove this result, we combine techniques that are used to find upper and lower bounds for the stable curve graph translation length with the Rauzy-Veech induction machinery.
We proceed with showing that for a fixed genus $g$ there is a sequence of pseudo-Anosovs $f_n$ with
$$\lim\limits_{n \to \infty} l_\mathcal{T}(f_n) = \infty \text{ and } l_\mathcal{C}(f_n) \le \frac{1}{g-1}$$ for all $n \in \mathbb{N}.$ As a corollary, we obtain that there are stable curve graph translation lengths with infinite multiplicity, i.e. there exists $q \in \mathbb{Q}$ and infinitely many, non-conjugate pseudo-Anosovs $f_n$ with $l_\mathcal{C}(f_n) = q$ for all $n.$
|
[
{
"version": "v1",
"created": "Mon, 27 Jan 2025 23:12:26 GMT"
},
{
"version": "v2",
"created": "Thu, 06 Feb 2025 22:22:54 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 22:00:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Bader",
"Philipp",
""
]
] |
2501.17057
|
Fernando G\'omez-Ortiz
|
Fernando G\'omez-Ortiz, Louis Bastogne, He Xu, Philippe Ghosez
|
Inhomogeneous Electric Fields for Precise Control and Displacement of Polar Textures
| null | null |
10.1103/PhysRevB.111.134107
| null |
cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
Since the discovery of polar topological textures, achieving efficient control and manipulation of them has emerged as a significant challenge for their integration into nanoelectronic devices. In this study, we use second principles molecular dynamic simulations to demonstrate the precise and reversible control of domain arrangements stabilizing diverse polarization textures through the application of various inhomogeneous electric fields. Furthermore, we conduct an in-depth study of ferroelectric domain motion under such fields, revealing features consistent with creep dynamics and establishing an upper limit for their propagation speed. Notably, our findings show that domain walls exhibit an asymmetric inertial response, present at the onset of the dynamics but absent during their cessation. These findings provide valuable insights into the dynamic behavior of polar textures, paving the way for the development of high-speed, low-power nanoelectronic applications.
|
[
{
"version": "v1",
"created": "Tue, 28 Jan 2025 16:33:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Gómez-Ortiz",
"Fernando",
""
],
[
"Bastogne",
"Louis",
""
],
[
"Xu",
"He",
""
],
[
"Ghosez",
"Philippe",
""
]
] |
2501.17082
|
Shuhan Jiang
|
Alberto S. Cattaneo, Shuhan Jiang
|
Equivariant localization in Batalin-Vilkovisky formalism
|
13 pages
|
Differential Geometry and its Applications, Volume 100, 2025, 102265
|
10.1016/j.difgeo.2025.102265
| null |
math-ph math.DG math.MP
|
http://creativecommons.org/licenses/by/4.0/
|
We derive equivariant localization formulas of Atiyah--Bott and cohomological field theory types in the Batalin-Vilkovisky formalism and discuss their applications in Poisson geometry and quantum field theory.
|
[
{
"version": "v1",
"created": "Tue, 28 Jan 2025 17:07:00 GMT"
},
{
"version": "v2",
"created": "Thu, 30 Jan 2025 09:21:57 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cattaneo",
"Alberto S.",
""
],
[
"Jiang",
"Shuhan",
""
]
] |
2501.18345
|
Smaranika Banerjee
|
Smaranika Banerjee, Anders Jerkstrand, Nigel Badnell, Quentin Pognan, Niamh Ferguson, Jon Grumer
|
Nebular spectra of kilonovae with detailed recombination rates -- I. Light r-process composition
|
Accepted to ApJ for publication
| null | null | null |
astro-ph.HE physics.atom-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
To investigate spectra of kilonovae in the NLTE phase (t>= 1 week), we perform atomic calculations for dielectronic recombination (DR) rates for the light r-process elements Se (Z = 34), Rb (Z = 37), Sr (Z = 38), Y (Z = 39), and Zr (Z = 40) using the HULLAC code. For the different elements, our results for the DR rate coefficients for recombining from the ionization states of II to I, III to II, and IV to III vary between 2x10^{-12} - 5x10^{-11} cm^3/s, 10^{-13} - 5x10^{-11} cm^3/s and 2x10^{-15} - 10^{-11} cm^3/s, respectively, at a temperature of T = 10,000 K. Using this new atomic data (DR), we study the impact on kilonova model spectra at phases of t = 10 days and t = 25 days after the merger using the spectral synthesis code SUMO. Compared to models using the previous treatment of recombination as a constant rate, the new models show significant changes in ionization and temperature, and correspondingly, in emergent spectra. With the new rates, we find that Zr (Z = 40) plays a yet more dominant role in kilonova spectra for light r-process compositions. Further, we show that previously predicted mid-infrared (e.g. [Se III] 4.55 mum) and optical (e.g. Rb I 7802, 7949 {\AA}) lines weaken in the new model. Instead [Se I] 5.03 mum emerges as a signature. These results demonstrate the importance of considering the detailed microphysics for modelling and interpreting the late-time kilonova spectra.
|
[
{
"version": "v1",
"created": "Thu, 30 Jan 2025 13:48:48 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Jul 2025 11:24:24 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 21:25:29 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Banerjee",
"Smaranika",
""
],
[
"Jerkstrand",
"Anders",
""
],
[
"Badnell",
"Nigel",
""
],
[
"Pognan",
"Quentin",
""
],
[
"Ferguson",
"Niamh",
""
],
[
"Grumer",
"Jon",
""
]
] |
2501.18844
|
Anthony Nicholas Zulli
|
Anthony N. Zulli, Brendan C. Mulkerin, Meera M. Parish, Jesper Levinsen
|
Universal Efimov Scaling in the Rabi-Coupled Few-Body Spectrum
|
16 pages, 10 figures + supplemental material
|
Phys. Rev. Lett. 135, 053401 (2025)
|
10.1103/yr6z-mtyq
| null |
cond-mat.quant-gas
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We investigate the behavior of the Efimov effect -- a universal quantum few-body phenomenon -- in the presence of an external driving field. Specifically, we consider up to three bosonic atoms, such as $^{133}$Cs, interacting with a light atom, such as $^{6}$Li, where the latter has two internal spin states $\{\uparrow, \downarrow\}$ that are Rabi coupled. Assuming that only the spin-$\uparrow$ light atom interacts with the bosons, we find that the Rabi drive transposes the entire Efimov spectrum such that the Efimov trimers and tetramers are centered around the Rabi-shifted two-body scattering resonance. Crucially, we show that the Rabi drive preserves the trimers' discrete scaling symmetry, while universally shifting the Efimov three-body parameter, leading to a log-periodic modulation in the spectrum as the Rabi drive is varied. Our results suggest that Efimov physics can be conveniently explored using an applied driving field, opening up the prospect of an externally tunable three-body parameter.
|
[
{
"version": "v1",
"created": "Fri, 31 Jan 2025 01:38:31 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 00:13:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zulli",
"Anthony N.",
""
],
[
"Mulkerin",
"Brendan C.",
""
],
[
"Parish",
"Meera M.",
""
],
[
"Levinsen",
"Jesper",
""
]
] |
2501.19280
|
Fabio Caccioli
|
Barak Budnick, Preben Forer, Pierpaolo Vivo, Sabrina Aufiero, Silvia Bartolucci, Fabio Caccioli
|
Top eigenpair statistics of diluted Wishart matrices
|
34 pages, 5 figures, accepted for publication in Journal of Physics A
| null |
10.1088/1751-8121/add821
| null |
cond-mat.stat-mech cond-mat.dis-nn math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Using the replica method, we compute the statistics of the top eigenpair of diluted covariance matrices of the form $\mathbf{J} = \mathbf{X}^T \mathbf{X}$, where $\mathbf{X}$ is a $N\times M$ sparse data matrix, in the limit of large $N,M$ with fixed ratio and a bounded number of nonzero entries. We allow for random non-zero weights, provided they lead to an isolated largest eigenvalue. By formulating the problem as the optimisation of a quadratic Hamiltonian constrained to the $N$-sphere at low temperatures, we derive a set of recursive distributional equations for auxiliary probability density functions, which can be efficiently solved using a population dynamics algorithm. The average largest eigenvalue is identified with a Lagrange parameter that governs the convergence of the algorithm, and the resulting stable populations are then used to evaluate the density of the top eigenvector's components. We find excellent agreement between our analytical results and numerical results obtained from direct diagonalisation.
|
[
{
"version": "v1",
"created": "Fri, 31 Jan 2025 16:43:28 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:44:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Budnick",
"Barak",
""
],
[
"Forer",
"Preben",
""
],
[
"Vivo",
"Pierpaolo",
""
],
[
"Aufiero",
"Sabrina",
""
],
[
"Bartolucci",
"Silvia",
""
],
[
"Caccioli",
"Fabio",
""
]
] |
2502.00424
|
Xinsheng Wang
|
Jian Li, Xinsheng Wang
|
Dense uniform Li-Yorke chaos for linear operators on a Banach space
|
16 pages, Accepted by International Journal of Bifurcation and Chaos
|
Internat. J. Bifur. Chaos Appl. Sci. Engrg. 35 (2025), no. 2, Paper No. 2550015, 10 pp
|
10.1142/S0218127425500154
| null |
math.DS
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
This paper focuses on the dense uniform Li-Yorke chaos for linear operators on a Banach space. Some sufficient conditions and equivalent conditions are established under which the dynamical system is densely uniformly Li-Yorke chaotic. It is shown that there are plenty of densely uniformly Li-Yorke chaotic operators. For unilateral backward weighted shifts and bilateral backward weighted shifts on $\ell^p$, it is shown that Li-Yorke chaos is equivalent to dense uniform Li-Yorke chaos.
|
[
{
"version": "v1",
"created": "Sat, 01 Feb 2025 13:02:25 GMT"
},
{
"version": "v2",
"created": "Wed, 05 Feb 2025 03:37:03 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Li",
"Jian",
""
],
[
"Wang",
"Xinsheng",
""
]
] |
2502.01582
|
Surajit Bera
|
Surajit Bera, Marco Schir\`o
|
Non-Stabilizerness of Sachdev-Ye-Kitaev Model
|
(15+10) Pages, (4+4) Figures, including Appendices
| null | null | null |
quant-ph cond-mat.dis-nn cond-mat.quant-gas cond-mat.stat-mech hep-th
|
http://creativecommons.org/licenses/by/4.0/
|
We study the non-stabilizerness or quantum magic of the Sachdev-Ye-Kitaev ($\rm SYK$) model, a prototype example of maximally chaotic quantum matter. We show that the Majorana spectrum of its ground state, encoding the spreading of the state in the Majorana basis, displays a Gaussian distribution as expected for chaotic quantum many-body systems. We compare our results with the case of the $\rm SYK_2$ model, describing non-chaotic random free fermions, and show that the Majorana spectrum is qualitatively different in the two cases, featuring an exponential Laplace distribution for the $\rm SYK_2$ model rather than a Gaussian. From the spectrum we extract the Stabilizer Renyi Entropy (SRE) and show that for both models it displays a linear scaling with system size, with a prefactor that is larger for the SYK model, which has therefore higher magic. Finally, we discuss the spreading of quantun magic under unitary dynamics, as described by the evolution of the Majorana spectrum and the Stabilizer Renyi Entropy starting from a stabilizer state. We show that the SRE for the $\rm SYK_2$ model equilibrates rapidly, but that in the steady-state the interacting chaotic SYK model has more magic than the simple $\rm SYK_2$. Our results suggest that the Majorana spectrum is qualitatively distinct in chaotic and non-chaotic many-body systems.
|
[
{
"version": "v1",
"created": "Mon, 03 Feb 2025 18:07:39 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 10:01:29 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Bera",
"Surajit",
""
],
[
"Schirò",
"Marco",
""
]
] |
2502.02171
|
Mohamed Youssef
|
Mohamed Youssef, Jian Peng, Oliver Bimber
|
DeepForest: Sensing Into Self-Occluding Volumes of Vegetation With Aerial Imaging
| null | null | null | null |
cs.CV eess.IV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Access to below-canopy volumetric vegetation data is crucial for understanding ecosystem dynamics. We address the long-standing limitation of remote sensing to penetrate deep into dense canopy layers. LiDAR and radar are currently considered the primary options for measuring 3D vegetation structures, while cameras can only extract the reflectance and depth of top layers. Using conventional, high-resolution aerial images, our approach allows sensing deep into self-occluding vegetation volumes, such as forests. It is similar in spirit to the imaging process of wide-field microscopy, but can handle much larger scales and strong occlusion. We scan focal stacks by synthetic-aperture imaging with drones and reduce out-of-focus signal contributions using pre-trained 3D convolutional neural networks with mean squared error (MSE) as the loss function. The resulting volumetric reflectance stacks contain low-frequency representations of the vegetation volume. Combining multiple reflectance stacks from various spectral channels provides insights into plant health, growth, and environmental conditions throughout the entire vegetation volume. Compared with simulated ground truth, our correction leads to ~x7 average improvements (min: ~x2, max: ~x12) for forest densities of 220 trees/ha - 1680 trees/ha. In our field experiment, we achieved an MSE of 0.05 when comparing with the top-vegetation layer that was measured with classical multispectral aerial imaging.
|
[
{
"version": "v1",
"created": "Tue, 04 Feb 2025 09:45:49 GMT"
},
{
"version": "v2",
"created": "Tue, 20 May 2025 12:05:04 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 10:02:07 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Youssef",
"Mohamed",
""
],
[
"Peng",
"Jian",
""
],
[
"Bimber",
"Oliver",
""
]
] |
2502.03093
|
Barbara Jasser
|
Barbara Jasser, Jovan Odavic and Alioscia Hamma
|
Stabilizer Entropy and entanglement complexity in the Sachdev-Ye-Kitaev model
| null | null | null | null |
quant-ph cond-mat.str-el hep-th nlin.CD nlin.SI
|
http://creativecommons.org/licenses/by/4.0/
|
The Sachdev-Ye-Kitaev (SYK) model is of paramount importance for the understanding of both strange metals and a microscopic theory of two-dimensional gravity. We study the interplay between Stabilizer R\'enyi Entropy (SRE) and entanglement entropy in both the ground state and highly excited states of the SYK4+SYK2 model interpolating the highly chaotic four-body interactions model with the integrable two-body interactions one. The interplay between these quantities is assessed also through universal statistics of the entanglement spectrum and its anti-flatness. We find that SYK4 is indeed characterized by a complex pattern of both entanglement and non-stabilizer resources while SYK2 is non-universal and not complex. We discuss the fragility and robustness of these features depending on the interpolation parameter.
|
[
{
"version": "v1",
"created": "Wed, 05 Feb 2025 11:32:59 GMT"
},
{
"version": "v2",
"created": "Sun, 23 Feb 2025 20:43:32 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 08:59:46 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Jasser",
"Barbara",
""
],
[
"Odavic",
"Jovan",
""
],
[
"Hamma",
"Alioscia",
""
]
] |
2502.05941
|
Sarthak Duary
|
Sarthak Duary, Pabitra Ray
|
Wave-form, memory and classical soft scalar theorems
|
23 pages, 1 figure, typos corrected, minor revision (physical reason for the absence of a temporal component of the trajectory coefficients in the probe approximation added, errors corrected in the earlier version for the resummed frequency space scalar wave-form, scalar wave-form in purely elastic two-body scattering subsection added), references added
| null |
10.1016/j.nuclphysb.2025.117044
| null |
hep-th
|
http://creativecommons.org/licenses/by-sa/4.0/
|
We examine a scattering process in which a set of particles come together, interact through a long-range massless scalar force and then disperse. We compute the trajectories of the scattered particles and derive an infinite series of subleading terms in the late-time and early-time expansion of the scalar wave-form. We study the scalar wave-form, the scalar memory term and classical soft scalar theorems. We ignore gravitational interactions and discuss the challenges for gravity.
|
[
{
"version": "v1",
"created": "Sun, 09 Feb 2025 15:56:05 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 12:20:38 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Duary",
"Sarthak",
""
],
[
"Ray",
"Pabitra",
""
]
] |
2502.06210
|
Aojun Lu
|
Aojun Lu, Junchao Ke, Chunhui Ding, Jiahao Fan, Jiancheng Lv and Yanan Sun
|
Achieving Deep Continual Learning via Evolution
| null | null | null | null |
cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Deep neural networks, despite their remarkable success, remain fundamentally limited in their ability to perform Continual Learning (CL). While most current methods aim to enhance the capabilities of a single model, Inspired by the collective learning mechanisms of human populations, we introduce Evolving Continual Learning (ECL), a framework that maintains and evolves a diverse population of neural network models. ECL continually searches for an optimal architecture for each introduced incremental task. This tailored model is trained on the corresponding task and archived as a specialized expert, contributing to a growing collection of skills. This approach inherently resolves the core CL challenges: stability is achieved through the isolation of expert models, while plasticity is greatly enhanced by evolving unique, task-specific architectures. Experimental results demonstrate that ECL significantly outperforms state-of-the-art individual-level CL methods. By shifting the focus from individual adaptation to collective evolution, ECL presents a novel path toward AI systems capable of CL.
|
[
{
"version": "v1",
"created": "Mon, 10 Feb 2025 07:21:44 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 03:59:49 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Lu",
"Aojun",
""
],
[
"Ke",
"Junchao",
""
],
[
"Ding",
"Chunhui",
""
],
[
"Fan",
"Jiahao",
""
],
[
"Lv",
"Jiancheng",
""
],
[
"Sun",
"Yanan",
""
]
] |
2502.06384
|
Anthony Parr
|
Ian Marquette and Anthony Parr
|
Polynomially Superintegrable Hamiltonians Separating in Cartesian Coordinates
|
43 pages
| null | null | null |
math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The problem of finding superintegrable Hamiltonians and their integrals of motion can be reduced to solving a series of compatibility equations that result from the overdetermination of the commutator or Poisson bracket relations. The computation of the compatibility equations requires a general formula for the coefficients, which in turn must depend on the potential to be solved for. This is in general a nonlinear problem and quite difficult. Thus, research has focused on dividing the classes of potential into standard and exotic ones so that a number of parameters may be set to zero and the coefficients may be obtained in a simpler setting. We have developed a new method in both the classical and quantum setting which readily yields a formula for the coefficients of the invariant without recourse to this division in the case of Cartesian-separable Hamiltonians. Even though they allow separation of variables as they in general involve potential in terms of higher transcendental and beyond hypergeometric for their wavefunctions, they are quite non-trivial models. The expressions we obtain are in general non-polynomial in the momenta whose fractional terms can be arbitrarily set to zero. These conditions are equivalent to the compatibility equations, but the only unknowns in addition to the potential are constant parameters. We also give the fourth-order standard potentials, and conjectures about general families.
|
[
{
"version": "v1",
"created": "Mon, 10 Feb 2025 12:12:56 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Feb 2025 01:49:56 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 13:18:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Marquette",
"Ian",
""
],
[
"Parr",
"Anthony",
""
]
] |
2502.06444
|
Xiaokang Du
|
Xiaokang Du, Huiling Liu, Qin Chang
|
Interpretation of 95 GeV Excess within the Georgi-Machacek Model in Light of Positive Definiteness Constraints
|
35 pages, 3 figures, 1 table; version published in Phys. Rev. D
|
Phys.Rev.D 112 (2025) 1, 015019
|
10.1103/w2xw-q49n
| null |
hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The recent observation of a di-photon excess around 95 GeV by the CMS and ATLAS Collaborations, along with the $b\bar{b}$ excess reported by the LEP Collaboration in the same mass region, has drawn significant interest in the possibility of new physics beyond the Standard Model (SM). The Georgi-Machacek (GM) model, which extends the Higgs sector of the SM by introducing additional triplet scalars while preserving custodial symmetry at tree level, provides a compelling framework to explain both excesses simultaneously via a light custodial singlet Higgs. In this work, we investigate whether the GM model can still accommodate these excesses when taking into account newly proposed vacuum stability constraints, particularly the positive definiteness conditions. Our numerical analysis not only confirms the existence of a viable parameter space capable of explaining the 95 GeV excesses, but also demonstrates that, compared to traditional tree-level constraints at the electroweak scale, the positive definiteness conditions further expand the allowed parameter space, thereby enhancing the viability of the GM model. Furthermore, we emphasize that future collider experiments will play a crucial role in testing this interpretation by refining Higgs coupling measurements and searching for additional Higgs bosons.
|
[
{
"version": "v1",
"created": "Mon, 10 Feb 2025 13:20:03 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 11:07:58 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Du",
"Xiaokang",
""
],
[
"Liu",
"Huiling",
""
],
[
"Chang",
"Qin",
""
]
] |
2502.08338
|
Christopher Barker
|
Christopher E. A. Barker, Charles Parton-Barr, Christopher H. Marrows, Olga Kazakova, and Craig Barton
|
Skyrmion motion in a synthetic antiferromagnet driven by asymmetric spin wave emission
| null | null | null | null |
cond-mat.mes-hall
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Skyrmions have been proposed as new information carriers in racetrack memory devices. To realise such devices, a small size; high speed of propagation; and minimal skyrmion Hall angle are required. Synthetic antiferromagnets (SAFs) present the ideal materials system to realise these aims. In this work, we use micromagnetic simulations to propose a new method for manipulating them using exclusively global magnetic fields. An out-of-plane microwave field induces oscillations in the skyrmions radius which in turn emits spin waves. When a static in-plane field is added, this breaks the symmetry of the skyrmions and causes asymmetric spin wave emission. This in turn drives motion of the skyrmions, with the fastest velocities observed at the frequency of the intrinsic out-of-phase breathing mode of the pair of skyrmions. This behaviour is investigated over a range of experimentally realistic antiferromagnetic interlayer exchange coupling strengths, and the results compared to previous works studying similar motion driven with an oscillating electric field. Through this the true effect of varying the exchange coupling strength is determined, and greater insight is gained into the mechanism of skyrmion motion. These results will help to inform the design of future novel computing architectures based on the dynamics of skyrmions in synthetic antiferromagnets.
|
[
{
"version": "v1",
"created": "Wed, 12 Feb 2025 12:05:20 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:04:02 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Barker",
"Christopher E. A.",
""
],
[
"Parton-Barr",
"Charles",
""
],
[
"Marrows",
"Christopher H.",
""
],
[
"Kazakova",
"Olga",
""
],
[
"Barton",
"Craig",
""
]
] |
2502.11641
|
Mladen Kova\v{c}evi\'c
|
Mladen Kova\v{c}evi\'c, Tatjana Grbi\'c, Darko \v{C}apko, Nemanja Nedi\'c, Srdjan Vukmirovi\'c
|
A Zero-Knowledge Proof for the Syndrome Decoding Problem in the Lee Metric
| null |
Transactions on Mathematical Cryptology, vol. 5, no. 1, 2025
| null | null |
cs.CR cs.IT math.IT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The syndrome decoding problem is one of the NP-complete problems lying at the foundation of code-based cryptography. The variant thereof where the distance between vectors is measured with respect to the Lee metric, rather than the more commonly used Hamming metric, has been analyzed recently in several works due to its potential relevance for building more efficient code-based cryptosystems. The purpose of this article is to present a zero-knowledge proof of knowledge for this variant of the problem.
|
[
{
"version": "v1",
"created": "Mon, 17 Feb 2025 10:35:18 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Feb 2025 10:38:21 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 12:46:58 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kovačević",
"Mladen",
""
],
[
"Grbić",
"Tatjana",
""
],
[
"Čapko",
"Darko",
""
],
[
"Nedić",
"Nemanja",
""
],
[
"Vukmirović",
"Srdjan",
""
]
] |
2502.11793
|
Roman Dubrovin
|
R.M. Dubrovin, A.V. Kimel, A.K. Zvezdin
|
Competition between terahertz magnetoelectric and N\'eel spin-orbit torque driven spin dynamics in metallic antiferromagnets
|
10 pages, 4 figures
|
Rhys. Rev. B (2025)
|
10.1103/srqk-m8rx
| null |
cond-mat.mtrl-sci
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Although magnetoelectric effects in metals are usually neglected, assuming that applied electric fields are screened by free charge carriers, the skin depth, defining the penetration depth of the fields, is non-zero and for THz electric fields typically reaches 400 nm. Hence, if the thickness of an antiferromagnetic film is of the order of tens of nm, electric field induced effects cannot be neglected. Here, we theoretically study the THz electric field induced spin dynamics in the metallic antiferromagnet $\mathrm{Mn}_{2}\mathrm{Au}$, whose spin arrangements allow it to exhibit a linear magnetoelectric effect. We show that the THz magnetoelectric torque in metallic antiferromagnets is proportional to the time derivative of the polarization induced by the THz electric field. Our simulations reveal that the magnetoelectric driven spin dynamics is indeed not negligible, and for a fair explanation of previously published experimental results in $\mathrm{Mn}_{2}\mathrm{Au}$ competition between THz magnetoelectric and N\'eel spin-orbit torques must be taken into account. Thus, it is shown that even in metallic antiferromagnets the THz magnetoelectric effect on spins can be strong and thus cannot be neglected.
|
[
{
"version": "v1",
"created": "Mon, 17 Feb 2025 13:33:17 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 08:30:06 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Dubrovin",
"R. M.",
""
],
[
"Kimel",
"A. V.",
""
],
[
"Zvezdin",
"A. K.",
""
]
] |
2502.12045
|
Tania Varesano
|
T. Varesano and D. M. Hassler and N. Zambrana Prado and J. M. Laming and J. Plowman and D.J. Knipp and M. Molnar and K. Barczynski and The SPICE consortium
|
FIP Bias Evolution in an Emerging Active Region as observed in SPICE Synoptic Observations
| null | null | null | null |
astro-ph.SR
|
http://creativecommons.org/licenses/by/4.0/
|
We investigate the time evolution of relative elemental abundances in the context of the first ionization potential effect focusing on an active region. Our aim is to characterize this evolution in different types of solar active region structures as well as in different atmospheric layers. We wish to assert how the measured changes relate to different magnetic topologies by computing abundance enhancement in different conditions using the ponderomotive force model.
Leveraging spectroscopic observations from the Spectral Imaging of the Coronal Environment instrument on board Solar Orbiter, we use extreme ultraviolet lines from ions formed across a broad temperature range--from the upper chromosphere to the low corona--and we perform relative abundance ratios following differential emission measure analysis. This methodology yields relative abundance maps from low, intermediate, and high first ionization potential elements. We obtain the temporal evolution of a number of abundance ratios for different structures on the Sun. We compare these results with the outcomes of the ponderomotive force model.
We find good correlation between the model and our results, suggesting an Alfv\'en-wave driven fractionation of the plasma. Fan loops, loop footpoints and active region boundaries exhibit coronal abundances, while the active region core shows more photospheric-like composition. A slow and steady increase in the magnesium to neon relative first ionization potential bias values is observed, starting around 1.5 and increasing by about 50\% after two days. The sulfur to oxygen evolution coupled with the model brings evidence of resonant waves fractionating the plasma in transition region structures.
|
[
{
"version": "v1",
"created": "Mon, 17 Feb 2025 17:12:37 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 17:49:34 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Varesano",
"T.",
""
],
[
"Hassler",
"D. M.",
""
],
[
"Prado",
"N. Zambrana",
""
],
[
"Laming",
"J. M.",
""
],
[
"Plowman",
"J.",
""
],
[
"Knipp",
"D. J.",
""
],
[
"Molnar",
"M.",
""
],
[
"Barczynski",
"K.",
""
],
[
"consortium",
"The SPICE",
""
]
] |
2502.12242
|
Ivica Smoli\'c
|
Jan Draga\v{s}evi\'c, Ina Moslavac and Ivica Smoli\'c
|
Weighing the curvature invariants
|
25 pages; ver. 3: published version (minor typos corrected)
|
Eur. Phys. J. C, 85 (2025) 818
|
10.1140/epjc/s10052-025-14552-9
|
ZTF-EP-25-04
|
gr-qc hep-th math-ph math.MP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We prove several inequalities between the curvature invariants, which impose constraints on curvature singularities. Some of the inequalities hold for a family of spacetimes which include static, Friedmann--Lema\^itre--Robertson--Walker, and Bianchi type I metrics, independently of whether they are solutions of some particular field equations. In contrast, others hold for solutions of Einstein's gravitational field equation and a family of energy-momentum tensors (featuring ideal fluids, scalar fields and nonlinear electromagnetic fields), independently of the specific form of the spacetime metric. We illustrate different behaviour of the basic curvature invariants with numerous examples and discuss the consequences and limitations of the proven results.
|
[
{
"version": "v1",
"created": "Mon, 17 Feb 2025 19:00:02 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Apr 2025 20:46:06 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 23:15:32 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Dragašević",
"Jan",
""
],
[
"Moslavac",
"Ina",
""
],
[
"Smolić",
"Ivica",
""
]
] |
2502.12266
|
Shan-Ming Ruan
|
Masamichi Miyaji, Shan-Ming Ruan, Shono Shibuya, Kazuyoshi Yano
|
Non-perturbative Overlaps in JT Gravity: From Spectral Form Factor to Generating Functions of Complexity
|
68 pages, many figures; v2: references added; v3: corrected typos, added a new plot, published version
|
JHEP 06 (2025) 251
|
10.1007/JHEP06(2025)251
|
YITP-25-28
|
hep-th gr-qc
|
http://creativecommons.org/licenses/by/4.0/
|
The interplay between black hole interior dynamics and quantum chaos provides a crucial framework for probing quantum effects in quantum gravity. In this work, we investigate non-perturbative overlaps in Jackiw-Teitelboim (JT) gravity to uncover universal signatures of quantum chaos and quantum complexity. Taking advantage of universal spectral correlators from random matrix theory, we compute the overlaps between the thermofield double (TFD) state and two distinct classes of states: fixed-length states, which encode maximal volume slices, and time-shifted TFD states. The squared overlaps naturally define probability distributions that quantify the expectation values of gravitational observables. Central to our results is the introduction of generating functions for quantum complexity measures, such as $\langle e^{-\alpha \ell} \rangle$. The time evolution of these generating functions exhibits the universal slope-ramp-plateau structure, mirroring the behavior of the spectral form factor (SFF). Using generating functions, we further demonstrate that the universal time evolution of complexity for chaotic systems, which is characterized by a linear growth followed by a late-time plateau, arises from the disappearance of the linear ramp as the regularization parameter $\alpha$ decreases. With regard to the time-shifted TFD state, we derive a surprising result: the expectation value of the time shift, which classically grows linearly, vanishes when non-perturbative quantum corrections are incorporated. This cancellation highlights a fundamental distinction between semiclassical and quantum gravitational descriptions of the black hole interior. All our findings establish generating functions as powerful probes of quantum complexity and chaos in gravitational and quantum systems.
|
[
{
"version": "v1",
"created": "Mon, 17 Feb 2025 19:06:30 GMT"
},
{
"version": "v2",
"created": "Wed, 12 Mar 2025 23:38:43 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 17:03:49 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Miyaji",
"Masamichi",
""
],
[
"Ruan",
"Shan-Ming",
""
],
[
"Shibuya",
"Shono",
""
],
[
"Yano",
"Kazuyoshi",
""
]
] |
2502.12831
|
Philibert Courau
|
Philibert Courau, Amaury Lambert, Emmanuel Schertzer
|
The gene's eye-view of quantitative genetics
|
(40 pages, 2 figures)
| null | null | null |
math.PR q-bio.PE
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Modelling the evolution of a continuous trait in a biological population is one of the oldest problems in evolutionary biology, which led to the birth of quantitative genetics. With the recent development of GWAS methods, it has become essential to link the evolution of the trait distribution to the underlying evolution of allelic frequencies at many loci, co-contributing to the trait value. The way most articles go about this is to make assumptions on the trait distribution, and use Wright's formula to model how the evolution of the trait translates on each individual locus. Here, we take a gene's eye-view of the system, starting from an explicit finite-loci model with selection, drift, recombination and mutation, in which the trait value is a direct product of the genome. We let the number of loci go to infinity under the assumption of strong recombination, and characterize the limit behavior of a given locus with a McKean-Vlasov SDE and the corresponding Fokker-Planck IPDE. In words, the selection on a typical locus depends on the mean behaviour of the other loci which can be approximated with the law of the focal locus. Results include the independence of two loci and explicit stationary distribution for allelic frequencies at a given locus (under some assumptions on the fitness function).
|
[
{
"version": "v1",
"created": "Tue, 18 Feb 2025 12:52:53 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:31:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Courau",
"Philibert",
""
],
[
"Lambert",
"Amaury",
""
],
[
"Schertzer",
"Emmanuel",
""
]
] |
2502.13698
|
Ella Orme
|
Ella S. C. Orme, Theodoulos Rodosthenous, Marina Evangelou
|
Multi-view biclustering via non-negative matrix tri-factorisation
| null | null | null | null |
stat.ME stat.AP
|
http://creativecommons.org/licenses/by/4.0/
|
Multi-view data is ever more apparent as methods for production, collection and storage of data become more feasible both practically and fiscally. However, not all features are relevant to describe the patterns for all individuals. Multi-view biclustering aims to simultaneously cluster both rows and columns, discovering clusters of rows as well as their view-specific identifying features. A novel multi-view biclustering approach based on non-negative matrix factorisation is proposed named ResNMTF. Demonstrated through extensive experiments on both synthetic and real datasets, ResNMTF successfully identifies both overlapping and non-exhaustive biclusters, without pre-existing knowledge of the number of biclusters present, and is able to incorporate any combination of shared dimensions across views. Further, to address the lack of a suitable bicluster-specific intrinsic measure, the popular silhouette score is extended to the bisilhouette score. The bisilhouette score is demonstrated to align well with known extrinsic measures, and proves useful as a tool for hyperparameter tuning as well as visualisation.
|
[
{
"version": "v1",
"created": "Wed, 19 Feb 2025 13:13:36 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 12:38:46 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Orme",
"Ella S. C.",
""
],
[
"Rodosthenous",
"Theodoulos",
""
],
[
"Evangelou",
"Marina",
""
]
] |
2502.14027
|
Peter Denton
|
Peter B. Denton, Julia Gehrlein, and Chui-Fan Kong
|
Testing New Physics in Oscillations at a Neutrino Factory
|
29 pages, 23 figures, 1 table, comments welcome! v2: matches published version
|
Nucl.Phys.B 1018 (2025) 117040
|
10.1016/j.nuclphysb.2025.117040
| null |
hep-ph hep-ex
|
http://creativecommons.org/licenses/by/4.0/
|
A neutrino factory is a potential successor to the upcoming generation of neutrino oscillation experiments and a possible precursor to next-generation muon colliders. Such a machine would provide a well-characterized beam of $\nu_\mu$, $\bar\nu_\mu$, $\nu_e$, and $\bar\nu_e$ neutrinos with comparable statistics. Here we show the sensitivity of a neutrino factory to new oscillation physics scenarios such as vector neutrino non-standard interactions and CPT violation. We study two different potential setups for a neutrino factory with different assumptions on charge identification in the far detector. We find that 10 years of a neutrino factory combined with 10 years of DUNE can improve over most of the current constraints on these scenarios and even over forecasted constraints by 20 years of DUNE. Additionally, we find that a neutrino factory can break degeneracies between the standard oscillation parameters and neutrino non-standard interaction parameters present at DUNE.
|
[
{
"version": "v1",
"created": "Wed, 19 Feb 2025 19:00:00 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:43:37 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Denton",
"Peter B.",
""
],
[
"Gehrlein",
"Julia",
""
],
[
"Kong",
"Chui-Fan",
""
]
] |
2502.15215
|
Seokhun Park
|
Seokhun Park, Insung Kong, Yongchan Choi, Chanmoo Park, Yongdai Kim
|
Tensor Product Neural Networks for Functional ANOVA Model
|
45 pages
| null | null | null |
stat.ML cs.LG math.ST stat.TH
|
http://creativecommons.org/licenses/by/4.0/
|
Interpretability for machine learning models is becoming more and more important as machine learning models become more complex. The functional ANOVA model, which decomposes a high-dimensional function into a sum of lower dimensional functions (commonly referred to as components), is one of the most popular tools for interpretable AI, and recently, various neural networks have been developed for estimating each component in the functional ANOVA model. However, such neural networks are highly unstable when estimating each component since the components themselves are not uniquely defined. That is, there are multiple functional ANOVA decompositions for a given function. In this paper, we propose a novel neural network which guarantees a unique functional ANOVA decomposition and thus is able to estimate each component stably and accurately. We call our proposed neural network ANOVA Tensor Product Neural Network (ANOVA-TPNN) since it is motivated by the tensor product basis expansion. Theoretically, we prove that ANOVA-TPNN can approximate any smooth function well. Empirically, we show that ANOVA-TPNN provide much more stable estimation of each component and thus much more stable interpretation when training data and initial values of the model parameters vary than existing neural networks do.
|
[
{
"version": "v1",
"created": "Fri, 21 Feb 2025 05:15:38 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Feb 2025 02:53:24 GMT"
},
{
"version": "v3",
"created": "Fri, 28 Feb 2025 15:00:20 GMT"
},
{
"version": "v4",
"created": "Sun, 08 Jun 2025 14:00:22 GMT"
},
{
"version": "v5",
"created": "Thu, 31 Jul 2025 00:17:08 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Park",
"Seokhun",
""
],
[
"Kong",
"Insung",
""
],
[
"Choi",
"Yongchan",
""
],
[
"Park",
"Chanmoo",
""
],
[
"Kim",
"Yongdai",
""
]
] |
2502.15276
|
Joe Moeller
|
Aaron D. Ames, Joe Moeller, Paulo Tabuada
|
Categorical Lyapunov Theory I: Stability of Flows
|
31 pages
| null | null | null |
math.DS cs.SY eess.SY math.CT
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Lyapunov's theorem provides a fundamental characterization of the stability of dynamical systems. This paper presents a categorical framework for Lyapunov theory, generalizing stability analysis with Lyapunov functions categorically. Core to our approach is the set of axioms underlying a setting for stability, which give the necessary ingredients for ``doing Lyapunov theory'' in a category of interest. With these minimal assumptions, we define the stability of equilibria, formulate Lyapunov morphisms, and demonstrate that the existence of Lyapunov morphisms is necessary and sufficient for establishing the stability of flows. To illustrate these constructions, we show how classical notions of stability, e.g., for continuous and discrete time dynamical systems, are captured by this categorical framework for Lyapunov theory. Finally, to demonstrate the extensibility of our framework, we illustrate how enriched categories, e.g., Lawvere metric spaces, yield settings for stability enabling one to ``do Lyapunov theory'' in enriched categories.
|
[
{
"version": "v1",
"created": "Fri, 21 Feb 2025 08:04:57 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 23:31:39 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ames",
"Aaron D.",
""
],
[
"Moeller",
"Joe",
""
],
[
"Tabuada",
"Paulo",
""
]
] |
2502.15468
|
Fani Dosopoulou
|
Fani Dosopoulou and Joseph Silk
|
Multi-messenger detection of black hole binaries in dark matter spikes
|
Accepted by PRL
| null | null | null |
astro-ph.HE astro-ph.CO astro-ph.GA
|
http://creativecommons.org/publicdomain/zero/1.0/
|
We investigate the inspiral of a high mass-ratio black hole binary located in the nucleus of a galaxy, where the primary central black hole is surrounded by a dense dark matter spike formed through accretion during the black hole growth phase. Within this spike, dark matter undergoes strong self-annihilation, producing a compact source of $\gamma$-ray radiation that is highly sensitive to spike density, while the binary emits gravitational waves at frequencies detectable by LISA. As the inspiralling binary interacts with the surrounding dark matter particles, it alters the density of the spike, thereby influencing the $\gamma$-ray flux from dark matter annihilation. We demonstrate that the spike self-annihilation luminosity decreases by $10\%$ to $90\%$ of its initial value, depending on the initial density profile and binary mass ratio, as the binary sweeps through the LISA band. This presents a new opportunity to indirectly probe dark matter through multi-messenger observations of galactic nuclei.
|
[
{
"version": "v1",
"created": "Fri, 21 Feb 2025 13:47:56 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 17:53:15 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Dosopoulou",
"Fani",
""
],
[
"Silk",
"Joseph",
""
]
] |
2502.15544
|
Xiaoyu Liu
|
Xiaoyu Liu, Caio Fabio Oliveira da Silva, Azita Dabiri, Yihui Wang, and Bart De Schutter
|
Learning-based model predictive control for passenger-oriented train rescheduling with flexible train composition
|
14 pages, 14 figures, submitted to journal
| null | null | null |
eess.SY cs.SY
|
http://creativecommons.org/licenses/by/4.0/
|
This paper focuses on passenger-oriented real-time train rescheduling, considering flexible train composition and rolling stock circulation, by integrating learning-based and optimization-based approaches. A learning-based model predictive control (MPC) approach is developed for real-time train rescheduling with flexible train composition and rolling stock circulation to address time-varying passenger demands. In the proposed approach, the values of the integer variables are obtained by pre-trained long short-term memory (LSTM) networks, while the continuous variables are determined through nonlinear constrained optimization. The learning-based MPC approach enables us to jointly consider efficiency and constraint satisfaction by combining learning-based and optimization-based approaches. In order to reduce the number of integer variables, four presolve techniques are developed to prune a subset of integer decision variables. Numerical simulations based on real-life data from the Beijing urban rail transit system are conducted to illustrate the effectiveness of the developed learning-based MPC approach.
|
[
{
"version": "v1",
"created": "Fri, 21 Feb 2025 15:51:30 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:02:01 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Liu",
"Xiaoyu",
""
],
[
"da Silva",
"Caio Fabio Oliveira",
""
],
[
"Dabiri",
"Azita",
""
],
[
"Wang",
"Yihui",
""
],
[
"De Schutter",
"Bart",
""
]
] |
2502.16421
|
Kaibin Zhou
|
Kaibin Zhou, Kaifeng Huang, Hao Deng, Zelin Tao, Ziniu Liu, Lin Zhang, Shengjie Zhao
|
Learning from Rendering: Realistic and Controllable Extreme Rainy Image Synthesis for Autonomous Driving Simulation
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Autonomous driving simulators provide an effective and low-cost alternative for evaluating or enhancing visual perception models. However, the reliability of evaluation depends on the diversity and realism of the generated scenes. Extreme weather conditions, particularly extreme rainfalls, are rare and costly to capture in real-world settings. While simulated environments can help address this limitation, existing rainy image synthesizers often suffer from poor controllability over illumination and limited realism, which significantly undermines the effectiveness of the model evaluation. To that end, we propose a learning-from-rendering rainy image synthesizer, which combines the benefits of the realism of rendering-based methods and the controllability of learning-based methods. To validate the effectiveness of our extreme rainy image synthesizer on semantic segmentation task, we require a continuous set of well-labeled extreme rainy images. By integrating the proposed synthesizer with the CARLA driving simulator, we develop CARLARain an extreme rainy street scene simulator which can obtain paired rainy-clean images and labels under complex illumination conditions. Qualitative and quantitative experiments validate that CARLARain can effectively improve the accuracy of semantic segmentation models in extreme rainy scenes, with the models' accuracy (mIoU) improved by 5% - 8% on the synthetic dataset and significantly enhanced in real extreme rainy scenarios under complex illuminations. Our source code and datasets are available at https://github.com/kb824999404/CARLARain/.
|
[
{
"version": "v1",
"created": "Sun, 23 Feb 2025 03:28:50 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 16:08:34 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhou",
"Kaibin",
""
],
[
"Huang",
"Kaifeng",
""
],
[
"Deng",
"Hao",
""
],
[
"Tao",
"Zelin",
""
],
[
"Liu",
"Ziniu",
""
],
[
"Zhang",
"Lin",
""
],
[
"Zhao",
"Shengjie",
""
]
] |
2502.17264
|
Jiayun Wu
|
Konstantina Bairaktari, Jiayun Wu, Zhiwei Steven Wu
|
Kandinsky Conformal Prediction: Beyond Class- and Covariate-Conditional Coverage
| null | null | null | null |
cs.LG stat.ML
|
http://creativecommons.org/licenses/by/4.0/
|
Conformal prediction is a powerful distribution-free framework for constructing prediction sets with coverage guarantees. Classical methods, such as split conformal prediction, provide marginal coverage, ensuring that the prediction set contains the label of a random test point with a target probability. However, these guarantees may not hold uniformly across different subpopulations, leading to disparities in coverage. Prior work has explored coverage guarantees conditioned on events related to the covariates and label of the test point. We present Kandinsky conformal prediction, a framework that significantly expands the scope of conditional coverage guarantees. In contrast to Mondrian conformal prediction, which restricts its coverage guarantees to disjoint groups -- reminiscent of the rigid, structured grids of Piet Mondrian's art -- our framework flexibly handles overlapping and fractional group memberships defined jointly on covariates and labels, reflecting the layered, intersecting forms in Wassily Kandinsky's compositions. Our algorithm unifies and extends existing methods, encompassing covariate-based group conditional, class conditional, and Mondrian conformal prediction as special cases, while achieving a minimax-optimal high-probability conditional coverage bound. Finally, we demonstrate the practicality of our approach through empirical evaluation on real-world datasets.
|
[
{
"version": "v1",
"created": "Mon, 24 Feb 2025 15:46:18 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:15:50 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Bairaktari",
"Konstantina",
""
],
[
"Wu",
"Jiayun",
""
],
[
"Wu",
"Zhiwei Steven",
""
]
] |
2502.17280
|
Gianluca Janka
|
Gianluca Janka, Herman Ishchenko, Zaher Salman, Andreas Suter, Thomas Prokscha
|
Monitoring the tagging efficiency of the Low-Energy Muon beamline through background analysis: Insights into the long-term performance of ultrathin carbon foils
| null | null | null | null |
physics.ins-det hep-ex
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The efficient tagging of individual muons, along with the stability of the involved ultrathin carbon foil, is critical for ensuring fast, reliable and reproducible low-energy muon spin rotation (LE-$\mathrm{\mu}$SR) measurements. At the Paul Scherrer Institute's Low-Energy Muon (LEM) beamline, we developed a method to monitor the tagging efficiency of the beamline using routinely collected muon decay histograms. This method leverages background comparison before and after the arrival time of the muons at the sample to extract the tagging efficiency, eliminating the need for additional detectors or measurements. By analyzing data collected between 2018 and 2024, we confirm the method's reliability and validate its results using independent reference measurements. Furthermore, we establish a correlation between the tagging efficiency and foil thickness by investigating the impact of contamination and outgassing on the carbon foil, as well as the restoration effects of laser cleaning. The findings underscore the importance of monitoring the carbon foil's condition to maintain consistent beamline performance and reproducibility. This method offers a practical approach for detector efficiency monitoring, applicable to other beamlines employing similar setups.
|
[
{
"version": "v1",
"created": "Mon, 24 Feb 2025 16:10:19 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Feb 2025 07:24:28 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 09:35:29 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Janka",
"Gianluca",
""
],
[
"Ishchenko",
"Herman",
""
],
[
"Salman",
"Zaher",
""
],
[
"Suter",
"Andreas",
""
],
[
"Prokscha",
"Thomas",
""
]
] |
2502.17402
|
Karen Kang
|
Karen Kang, Simona J. Miller, Katerina Chatziioannou, Deborah Ferguson
|
Mapping Parameter Correlations in Spinning Binary Black Hole Mergers
|
21 pages including appendices and bibliography; 19 figures. Version accepted to PRD
| null |
10.1103/4vbp-1jwl
| null |
gr-qc astro-ph.HE
|
http://creativecommons.org/licenses/by/4.0/
|
The spins of binary black holes measured with gravitational waves provide insights about the formation, evolution, and dynamics of these systems. However, interpreting these measurements-especially for heavy black holes-remains an open problem. While the imprint of spin during the inspiral phase, where the black holes are well-separated, is understood through analytic descriptions of the dynamics, no such expressions exist for the merger. Though numerical relativity simulations provide an exact solution (to within numerical error), the imprint of the full six spin degrees of freedom on the signal is not transparent. In the absence of analytic expressions for the merger and to advance our ability to interpret massive binary black hole spin measurements, here we propose a waveform-based approach. Leveraging a neural network to efficiently calculate mismatches between waveforms, we identify regions in the parameter space of spins and mass ratio that result in low mismatches and thus similar waveforms. We map these regions with a Gaussian fit, thus identifying correlations and quantifying their strength. For low-mass, inspiral-dominated systems, we recover the known physical imprint: larger aligned spins are correlated with more equal masses as they have opposite effects on the inspiral length. For high-mass, merger-dominated signals, a qualitatively similar correlation is present, though its shape is altered and strength decreases with larger total mass. Correlations between in-plane spins and mass ratio follow a similar trend, with their shape and strength altered as the mass increases. Our new methodology of waveform-based correlation mapping provides a first step toward systematically modeling spin effects in merger-dominated signals across the full intrinsic parameter-space and motivates future effective spin parameters beyond the reach of analytic methods.
|
[
{
"version": "v1",
"created": "Mon, 24 Feb 2025 18:30:07 GMT"
},
{
"version": "v2",
"created": "Tue, 29 Jul 2025 16:58:01 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kang",
"Karen",
""
],
[
"Miller",
"Simona J.",
""
],
[
"Chatziioannou",
"Katerina",
""
],
[
"Ferguson",
"Deborah",
""
]
] |
2502.17482
|
Ziwei Wang
|
Ziwei Wang, Siyang Li, Xiaoqing Chen, and Dongrui Wu
|
MVCNet: Multi-View Contrastive Network for Motor Imagery Classification
|
12 pages, 9 figures
| null | null | null |
eess.SP cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Electroencephalography (EEG)-based brain-computer interfaces (BCIs) enable neural interaction by decoding brain activity for external communication. Motor imagery (MI) decoding has received significant attention due to its intuitive mechanism. However, most existing models rely on single-stream architectures and overlook the multi-view nature of EEG signals, leading to limited performance and generalization. We propose a multi-view contrastive network (MVCNet), a dual-branch architecture that parallelly integrates CNN and Transformer blocks to capture both local spatial-temporal features and global temporal dependencies. To enhance the informativeness of training data, MVCNet incorporates a unified augmentation pipeline across time, frequency, and spatial domains. Two contrastive modules are further introduced: a cross-view contrastive module that enforces consistency of original and augmented views, and a cross-model contrastive module that aligns features extracted from both branches. Final representations are fused and jointly optimized by contrastive and classification losses. Experiments on five public MI datasets across three scenarios demonstrate that MVCNet consistently outperforms nine state-of-the-art MI decoding networks, highlighting its effectiveness and generalization ability. MVCNet provides a robust solution for MI decoding by integrating multi-view information and dual-branch modeling, contributing to the development of more reliable BCI systems.
|
[
{
"version": "v1",
"created": "Tue, 18 Feb 2025 10:30:53 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Feb 2025 10:31:05 GMT"
},
{
"version": "v3",
"created": "Sun, 27 Apr 2025 19:58:19 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 08:10:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Ziwei",
""
],
[
"Li",
"Siyang",
""
],
[
"Chen",
"Xiaoqing",
""
],
[
"Wu",
"Dongrui",
""
]
] |
2502.17707
|
Mathieu Baillif
|
Mathieu Baillif
|
On non-Hausdorff manifolds
|
Work in progress whose contents might change over time. Some pictures use colors and can be difficult to decipher for color blind readers. V2: See the list of main changes on first page. Also, typos corrected and improved presentation in many places
| null | null | null |
math.GN
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
We investigate various aspects of non-Hausdorff manifolds (NH-manifold for short). First, we extend some known results about which covering properties, together with (a weakening of) homogeneity, imply that a manifold is Hausdorff. Writing $NH_X(x)$ for the subset of points of a space $X$ which cannot be separated of $x$ by open sets, we exhibit homogeneous NH-manifolds $M$ of dimension $\ge 2$ such that $NH_M(x)$ is non-homogeneous (in particular, non discrete), and give various properties that imply discreteness of $NH_M(x)$. We show that there is an everywhere non-Hausdorff hereditarily separable manifold under CH and that the existence of an everywhere non-Hausdorff hereditarily separable manifold in a particular class implies that of a locally compact S-space. We also construct various examples of somewhat pathological NH-manifolds. In particular, we show that there are NH-manifolds $M$ with a point $x\in M$ such that $NH_M(x)$ is a copy of: the Cantor space, any special tree, any $\Psi$-space (dimension $1$), $\mathbb{R}$, the long ray, (part of) boundaries of domains in $\mathbb{R}^2$ (dimension $2$). We also build a homogeneous $1$-dimensional NH-manifold $M$ such that for each $x,y$ there is some $z$ with $x\in NH_M(z)$, $z\in NH_M(y)$. We start by recalling basic properties of NH-manifolds and use elementary (or at least well known) methods of general or set theoretic topology, with a little bit of conformal theory for some of the examples. Many pictures are given to illustrate the constructions and the proofs are rather detailed, which is the main reason for the length of this note.
|
[
{
"version": "v1",
"created": "Mon, 24 Feb 2025 23:00:04 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:46:59 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Baillif",
"Mathieu",
""
]
] |
2502.18037
|
Carlo Andrea Rozzi
|
Carlo Andrea Rozzi, Annamaria Lisotti, Guido Goldoni, Valentina De Renzi
|
Listen! It's a phase transition. The sound of a shape-memory alloy
| null | null |
10.1119/5.0217522
| null |
cond-mat.mtrl-sci physics.ed-ph
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Shape-memory alloys exhibit a solid-to-solid phase transition that involves a temperature-driven rearrangement of their crystal structure and is responsible for their remarkable properties and numerous technological applications. Here, we propose a simple experiment that analyzes the sound emitted by a Ni$_{40}$Ti$_{50}$Cu$_{10}$ bar at different temperatures as it undergoes a transition between its austenite and martensite phases. We show that the phase transition, which occurs slightly above room temperature, can be qualitatively detected by the ear and quantitatively described using a very simple experimental setup and sound analysis tools. Such a sound-based investigation provides an unusual and engaging way to experimentally introduce solid-to-solid phase transitions, that is suitable for undergraduate courses.
|
[
{
"version": "v1",
"created": "Tue, 25 Feb 2025 09:52:51 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 12:04:26 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Rozzi",
"Carlo Andrea",
""
],
[
"Lisotti",
"Annamaria",
""
],
[
"Goldoni",
"Guido",
""
],
[
"De Renzi",
"Valentina",
""
]
] |
2502.18041
|
Yunpeng Gao
|
Yunpeng Gao, Chenhui Li, Zhongrui You, Junli Liu, Zhen Li, Pengan Chen, Qizhi Chen, Zhonghan Tang, Liansheng Wang, Penghui Yang, Yiwen Tang, Yuhang Tang, Shuai Liang, Songyi Zhu, Ziqin Xiong, Yifei Su, Xinyi Ye, Jianan Li, Yan Ding, Dong Wang, Zhigang Wang, Bin Zhao, Xuelong Li
|
OpenFly: A Comprehensive Platform for Aerial Vision-Language Navigation
|
20 pages, 11 figures
| null | null | null |
cs.CV cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Vision-Language Navigation (VLN) aims to guide agents by leveraging language instructions and visual cues, playing a pivotal role in embodied AI. Indoor VLN has been extensively studied, whereas outdoor aerial VLN remains underexplored. The potential reason is that outdoor aerial view encompasses vast areas, making data collection more challenging, which results in a lack of benchmarks. To address this problem, we propose OpenFly, a platform comprising various rendering engines, a versatile toolchain, and a large-scale benchmark for aerial VLN. Firstly, we integrate diverse rendering engines and advanced techniques for environment simulation, including Unreal Engine, GTA V, Google Earth, and 3D Gaussian Splatting (3D GS). Particularly, 3D GS supports real-to-sim rendering, further enhancing the realism of our environments. Secondly, we develop a highly automated toolchain for aerial VLN data collection, streamlining point cloud acquisition, scene semantic segmentation, flight trajectory creation, and instruction generation. Thirdly, based on the toolchain, we construct a large-scale aerial VLN dataset with 100k trajectories, covering diverse heights and lengths across 18 scenes. Moreover, we propose OpenFly-Agent, a keyframe-aware VLN model emphasizing key observations during flight. For benchmarking, extensive experiments and analyses are conducted, evaluating several recent VLN methods and showcasing the superiority of our OpenFly platform and agent. The toolchain, dataset, and codes will be open-sourced.
|
[
{
"version": "v1",
"created": "Tue, 25 Feb 2025 09:57:18 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Feb 2025 02:10:39 GMT"
},
{
"version": "v3",
"created": "Tue, 04 Mar 2025 08:38:58 GMT"
},
{
"version": "v4",
"created": "Sat, 08 Mar 2025 10:11:32 GMT"
},
{
"version": "v5",
"created": "Wed, 21 May 2025 12:40:06 GMT"
},
{
"version": "v6",
"created": "Thu, 31 Jul 2025 07:55:21 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Gao",
"Yunpeng",
""
],
[
"Li",
"Chenhui",
""
],
[
"You",
"Zhongrui",
""
],
[
"Liu",
"Junli",
""
],
[
"Li",
"Zhen",
""
],
[
"Chen",
"Pengan",
""
],
[
"Chen",
"Qizhi",
""
],
[
"Tang",
"Zhonghan",
""
],
[
"Wang",
"Liansheng",
""
],
[
"Yang",
"Penghui",
""
],
[
"Tang",
"Yiwen",
""
],
[
"Tang",
"Yuhang",
""
],
[
"Liang",
"Shuai",
""
],
[
"Zhu",
"Songyi",
""
],
[
"Xiong",
"Ziqin",
""
],
[
"Su",
"Yifei",
""
],
[
"Ye",
"Xinyi",
""
],
[
"Li",
"Jianan",
""
],
[
"Ding",
"Yan",
""
],
[
"Wang",
"Dong",
""
],
[
"Wang",
"Zhigang",
""
],
[
"Zhao",
"Bin",
""
],
[
"Li",
"Xuelong",
""
]
] |
2502.18566
|
Joshua Lockyer
|
Suchita Kulkarni, Joshua Lockyer, Matthew J. Strassler
|
On the Simulation of Hidden Parton Showers in the Conformal Window
|
24 pages, 8 figures, version accepted for publication
| null | null | null |
hep-ph hep-ex
|
http://creativecommons.org/licenses/by/4.0/
|
We consider confining Hidden Valley/Dark Sector theories containing many dark quark flavors. These theories are in the ``conformal window'': they reach an infrared fixed point when their quarks are massless, and have unfamiliar confinement when the quark masses are non-zero but small. Their jets of hidden hadrons may be quite different from those familiar from QCD, but their details cannot currently be simulated even qualitatively. This is partly due to the use of approximations to the two-loop running coupling in existing event generators' parton showers, which are not broadly applicable across the conformal window. We argue that the exact two-loop running coupling, and a corresponding Sudakov factor employing that coupling, must be implemented in simulation packages in order to allow phenomenological studies of these theories.
|
[
{
"version": "v1",
"created": "Tue, 25 Feb 2025 19:00:11 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:44:22 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kulkarni",
"Suchita",
""
],
[
"Lockyer",
"Joshua",
""
],
[
"Strassler",
"Matthew J.",
""
]
] |
2502.18788
|
Efstathios Konstantinos Chrontsios Garitsis
|
Efstathios Konstantinos Chrontsios Garitsis, Vyron Vellis
|
H\"older spiral arcs
|
11 pages, proved a more general theorem in Section 3
| null | null | null |
math.CA math.DS
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We establish a quantitative necessary and sufficient condition for a spiral arc to be a H\"older arc. The class of spiral arcs contains the polynomial spirals studied by Fraser, and the elliptical spirals studied by Burrell-Falconer-Fraser. As an application, we recover the sharp result on the H\"older winding problem for polynomial spirals. Moreover, we provide a sharp exponent estimate for the H\"older classification of polynomial spirals, which coincides with the corresponding quasiconformal classification estimate, and improve certain exponent bounds of Burrell-Falconer-Fraser on the H\"older classification of elliptical spirals.
|
[
{
"version": "v1",
"created": "Wed, 26 Feb 2025 03:45:11 GMT"
},
{
"version": "v2",
"created": "Sat, 12 Apr 2025 02:49:36 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Jul 2025 19:42:48 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Garitsis",
"Efstathios Konstantinos Chrontsios",
""
],
[
"Vellis",
"Vyron",
""
]
] |
2502.19299
|
Alexis Anagnostakis
|
Alexis Anagnostakis
|
General diffusions on the star graph as time-changed Walsh Brownian motion
| null | null | null | null |
math.PR
|
http://creativecommons.org/licenses/by/4.0/
|
We establish the representation of general regular diffusions on star-shaped graphs as time-changed Walsh Brownian motions. These are regular continuous Markov processes described locally by a family generalized second order differential operators defined on every edge and a gluing condition at the junction vertex. This allows us to prove two additional results: (i) A representation of diffusions with sticky gluing conditions as time-changes of diffusions governed by the same differential operators but with non-sticky gluing conditions. (ii) An occupation times formula for such diffusions, analogous to the classical It\^o--McKean formula for one-dimensional diffusions.
Additionally, we prove two results of independent interest. First, conditions under which a diffusion on the star graph is Feller and Feller--Dynkin, extending classical results for one-dimensional diffusions. Second, the existence uniqueness of solutions to the Dirichlet problem on the unit disk of the star graph for a general diffusion operator and explicit expressions for its solution.
|
[
{
"version": "v1",
"created": "Wed, 26 Feb 2025 16:56:58 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 14:37:47 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Anagnostakis",
"Alexis",
""
]
] |
2502.20263
|
Rongzhen Zhao
|
Rongzhen Zhao, Vivienne Wang, Juho Kannala, Joni Pajarinen
|
Vector-Quantized Vision Foundation Models for Object-Centric Learning
|
Accepted by ACM MM 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Perceiving visual scenes as objects and background--like humans do--Object-Centric Learning (OCL) aggregates image or video feature maps into object-level feature vectors, termed \textit{slots}. OCL's self-supervision of reconstructing the input from these aggregated slots struggles with complex object textures, thus Vision Foundation Model (VFM) representations are used as the aggregation input and reconstruction target. However, existing methods leverage VFM representations in diverse ways and often fail to fully exploit their potential. In response, we propose a clean architecture--Vector-Quantized VFMs for OCL (VQ-VFM-OCL, or VVO)--that unifies mainstream OCL methods. The key to our unification is simple yet effective, just shared quantizing the same VFM representation as the reconstruction target. Through mathematical modeling and statistical verification, we further analyze why VFM representations facilitate OCL aggregation and how their shared quantization as reconstruction targets strengthens OCL supervision. Experiments show that across different VFMs, aggregators and decoders, our VVO consistently outperforms baselines in object discovery and recognition, as well as downstream visual prediction and reasoning. The implementation and model checkpoints are available on https://github.com/Genera1Z/VQ-VFM-OCL.
|
[
{
"version": "v1",
"created": "Thu, 27 Feb 2025 16:51:13 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 08:41:06 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 08:40:53 GMT"
},
{
"version": "v4",
"created": "Thu, 31 Jul 2025 17:36:31 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhao",
"Rongzhen",
""
],
[
"Wang",
"Vivienne",
""
],
[
"Kannala",
"Juho",
""
],
[
"Pajarinen",
"Joni",
""
]
] |
2502.20632
|
Shoummo Khandoker
|
Shoummo Ahsan Khandoker, Estelle M. Inack, Mohamed Hibat-Allah
|
Lattice Protein Folding with Variational Annealing
| null | null | null | null |
cond-mat.dis-nn cs.AI cs.LG q-bio.BM
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Understanding the principles of protein folding is a cornerstone of computational biology, with implications for drug design, bioengineering, and the understanding of fundamental biological processes. Lattice protein folding models offer a simplified yet powerful framework for studying the complexities of protein folding, enabling the exploration of energetically optimal folds under constrained conditions. However, finding these optimal folds is a computationally challenging combinatorial optimization problem. In this work, we introduce a novel upper-bound training scheme that employs masking to identify the lowest-energy folds in two-dimensional Hydrophobic-Polar (HP) lattice protein folding. By leveraging Dilated Recurrent Neural Networks (RNNs) integrated with an annealing process driven by temperature-like fluctuations, our method accurately predicts optimal folds for benchmark systems of up to 60 beads. Our approach also effectively masks invalid folds from being sampled without compromising the autoregressive sampling properties of RNNs. This scheme is generalizable to three spatial dimensions and can be extended to lattice protein models with larger alphabets. Our findings emphasize the potential of advanced machine learning techniques in tackling complex protein folding problems and a broader class of constrained combinatorial optimization challenges.
|
[
{
"version": "v1",
"created": "Fri, 28 Feb 2025 01:30:15 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 19:46:48 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Khandoker",
"Shoummo Ahsan",
""
],
[
"Inack",
"Estelle M.",
""
],
[
"Hibat-Allah",
"Mohamed",
""
]
] |
2502.20760
|
Weijia Zhang
|
Weijia Zhang, Fei Xie, Weidong Cai, Chao Ma
|
VRM: Knowledge Distillation via Virtual Relation Matching
|
Accepted by ICCV 2025 (Highlight)
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Knowledge distillation (KD) aims to transfer the knowledge of a more capable yet cumbersome teacher model to a lightweight student model. In recent years, relation-based KD methods have fallen behind, as their instance-matching counterparts dominate in performance. In this paper, we revive relational KD by identifying and tackling several key issues in relation-based methods, including their susceptibility to overfitting and spurious responses. Specifically, we transfer novelly constructed affinity graphs that compactly encapsulate a wealth of beneficial inter-sample, inter-class, and inter-view correlations by exploiting virtual views and relations as a new kind of knowledge. As a result, the student has access to richer guidance signals and stronger regularisation throughout the distillation process. To further mitigate the adverse impact of spurious responses, we prune the affinity graphs by dynamically detaching redundant and unreliable edges. Extensive experiments on CIFAR-100, ImageNet, and MS-COCO datasets demonstrate the superior performance of the proposed virtual relation matching (VRM) method, where it consistently sets new state-of-the-art records over a range of models, architectures, tasks, and set-ups. For instance, VRM for the first time hits 74.0% accuracy for ResNet50-to-MobileNetV2 distillation on ImageNet, and improves DeiT-T by 14.44% on CIFAR-100 with a ResNet56 teacher.
|
[
{
"version": "v1",
"created": "Fri, 28 Feb 2025 06:29:39 GMT"
},
{
"version": "v2",
"created": "Tue, 01 Apr 2025 04:57:26 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 06:09:20 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhang",
"Weijia",
""
],
[
"Xie",
"Fei",
""
],
[
"Cai",
"Weidong",
""
],
[
"Ma",
"Chao",
""
]
] |
2502.20934
|
Utku Ozbulak
|
Utku Ozbulak, Seyed Amir Mousavi, Francesca Tozzi, Niki Rashidian, Wouter Willaert, Wesley De Neve, Joris Vankerschaver
|
Revisiting the Evaluation Bias Introduced by Frame Sampling Strategies in Surgical Video Segmentation Using SAM2
|
Accepted for publication in the 28th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) Workshop on Fairness of AI in Medical Imaging (FAIMI), 2025
| null | null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Real-time video segmentation is a promising opportunity for AI-assisted surgery, offering intraoperative guidance by identifying tools and anatomical structures. Despite growing interest in surgical video segmentation, annotation protocols vary widely across datasets -- some provide dense, frame-by-frame labels, while others rely on sparse annotations sampled at low frame rates such as 1 FPS. In this study, we investigate how such inconsistencies in annotation density and frame rate sampling influence the evaluation of zero-shot segmentation models, using SAM2 as a case study for cholecystectomy procedures. Surprisingly, we find that under conventional sparse evaluation settings, lower frame rates can appear to outperform higher ones due to a smoothing effect that conceals temporal inconsistencies. However, when assessed under real-time streaming conditions, higher frame rates yield superior segmentation stability, particularly for dynamic objects like surgical graspers. To understand how these differences align with human perception, we conducted a survey among surgeons, nurses, and machine learning engineers and found that participants consistently preferred high-FPS segmentation overlays, reinforcing the importance of evaluating every frame in real-time applications rather than relying on sparse sampling strategies. Our findings highlight the risk of evaluation bias that is introduced by inconsistent dataset protocols and bring attention to the need for temporally fair benchmarking in surgical video AI.
|
[
{
"version": "v1",
"created": "Fri, 28 Feb 2025 10:42:09 GMT"
},
{
"version": "v2",
"created": "Mon, 07 Apr 2025 13:22:10 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 02:07:41 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ozbulak",
"Utku",
""
],
[
"Mousavi",
"Seyed Amir",
""
],
[
"Tozzi",
"Francesca",
""
],
[
"Rashidian",
"Niki",
""
],
[
"Willaert",
"Wouter",
""
],
[
"De Neve",
"Wesley",
""
],
[
"Vankerschaver",
"Joris",
""
]
] |
2502.20962
|
Anshuman Tripathi
|
Anshuman Tripathi, Abhirup Datta, Aishrila Mazumder, Suman Majumdar
|
Impact of Calibration and Position Errors on Astrophysical Parameters of the HI 21cm Signal
|
27 pages, 7 figures, comments are welcome, prepared for submission to JCAP
| null | null | null |
astro-ph.CO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The Epoch of Reionization (EoR) and Cosmic Dawn (CD) are pivotal stages during the first billion years of the universe, exerting a significant influence on the development of cosmic structure. The detection of the redshifted 21-cm signal from these epochs is challenging due to the dominance of significantly stronger astrophysical foregrounds and the presence systematics. This work used the 21cm E2E (end to end) pipeline, followed by simulation methodology described \cite{2022Mazumder} to conduct synthetic observations of a simulated sky model that includes both the redshifted 21-cm signal and foregrounds. A framework was constructed using Artificial Neural Networks (ANN) and Bayesian techniques to directly deduce astrophysical parameters from the measured power spectrum. This approach eliminates the need for explicit telescope layout effects correction in interferometric arrays such as SKA-Low. The present work investigates the impact of gain calibration and sky model position errors on the recovery of the redshifted 21-cm power spectrum for the SKA-Low AA$^{\ast}$ array configuration. We assessed the effects of these inaccuracies on the deduced astrophysical parameters and established acceptable tolerance levels. Based on our results, the gain calibration error tolerance for ideal signal detection is 0.001 \%. However, if the sky model position errors exceed 0.048 arcseconds, the remaining foregrounds would obscure the target signal.
|
[
{
"version": "v1",
"created": "Fri, 28 Feb 2025 11:23:48 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 14:36:40 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Tripathi",
"Anshuman",
""
],
[
"Datta",
"Abhirup",
""
],
[
"Mazumder",
"Aishrila",
""
],
[
"Majumdar",
"Suman",
""
]
] |
2502.21066
|
Sebastjan Glinsek
|
Nagamalleswara Rao Alluri, Longfei Song, Stephanie Girod, Barnik Mandal, Juliette Cardoletti, Vid Bobnar, Torsten Granzow, Veronika Kovacova, Adrian-Marie Philippe, Emmanuel Defay, Sebastjan Glinsek
|
Solution-Based Fabrication of High-Performance K$_{0.5}$Na$_{0.5}$NbO$_3$ Thin Films for Surface Haptics
| null | null | null | null |
cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
K$_{0.5}$Na$_{0.5}$NbO$_3$ is among the most promising lead-free piezoelectrics. While its sputtered films match the performance of the champion piezoelectric Pb(Zr,Ti)O$_3$, reproducible processing of high-quality and time-stable solution-processed K$_{0.5}$Na$_{0.5}$NbO$_3$ films remains challenging. Here, we report 1 $\mu$m-thick Mn-doped K$_{0.5}$Na$_{0.5}$NbO$_3$ films prepared through a chemical solution deposition process, which have perfectly dense microstructure and uniform composition across their thickness. The films exhibit a high transverse piezoelectric coefficient ($e_{31,f} = -15.4$ C/m$^2$), high dielectric permittivity ($\varepsilon_r \approx 920$), low dielectric losses ($\tan\delta = 0.05$) and can withstand electric fields up to at least 1 MV/cm. The functional properties show excellent stability over time, and the synthesis process is reproducible. Furthermore, a surface acoustic haptic device is demonstrated by using K$_{0.5}$Na$_{0.5}$NbO$_3$ thin-film actuators. The results demonstrate the high potential of Mn-doped K$_{0.5}$Na$_{0.5}$NbO$_3$ films to become a replacement for lead-based Pb(Zr,Ti)O$_3$ films in piezoelectric applications.
|
[
{
"version": "v1",
"created": "Fri, 28 Feb 2025 14:02:39 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 05:45:33 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Alluri",
"Nagamalleswara Rao",
""
],
[
"Song",
"Longfei",
""
],
[
"Girod",
"Stephanie",
""
],
[
"Mandal",
"Barnik",
""
],
[
"Cardoletti",
"Juliette",
""
],
[
"Bobnar",
"Vid",
""
],
[
"Granzow",
"Torsten",
""
],
[
"Kovacova",
"Veronika",
""
],
[
"Philippe",
"Adrian-Marie",
""
],
[
"Defay",
"Emmanuel",
""
],
[
"Glinsek",
"Sebastjan",
""
]
] |
2503.01199
|
Kaimin Liao
|
Kaimin Liao, Hua Wang, Zhi Chen, Luchao Wang, Yaohua Tang
|
LiteGS: A High-performance Framework to Train 3DGS in Subminutes via System and Algorithm Codesign
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
3D Gaussian Splatting (3DGS) has emerged as promising alternative in 3D representation. However, it still suffers from high training cost. This paper introduces LiteGS, a high performance framework that systematically optimizes the 3DGS training pipeline from multiple aspects. At the low-level computation layer, we design a ``warp-based raster'' associated with two hardware-aware optimizations to significantly reduce gradient reduction overhead. At the mid-level data management layer, we introduce dynamic spatial sorting based on Morton coding to enable a performant ``Cluster-Cull-Compact'' pipeline and improve data locality, therefore reducing cache misses. At the top-level algorithm layer, we establish a new robust densification criterion based on the variance of the opacity gradient, paired with a more stable opacity control mechanism, to achieve more precise parameter growth. Experimental results demonstrate that LiteGS accelerates the original 3DGS training by up to 13.4x with comparable or superior quality and surpasses the current SOTA in lightweight models by up to 1.4x speedup. For high-quality reconstruction tasks, LiteGS sets a new accuracy record and decreases the training time by an order of magnitude.
|
[
{
"version": "v1",
"created": "Mon, 03 Mar 2025 05:52:02 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 07:35:04 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Liao",
"Kaimin",
""
],
[
"Wang",
"Hua",
""
],
[
"Chen",
"Zhi",
""
],
[
"Wang",
"Luchao",
""
],
[
"Tang",
"Yaohua",
""
]
] |
2503.02653
|
Xiaokang Du
|
LiangLiang Shang, Yuanping Wang, Xiaokang Du, Bingfang Yang, Stefano Moretti
|
Triplet fermion WIMP dark matter in the general supersymmetric Georgi-Machacek model
| null |
Phys.Rev.D 111 (2025) 11, 115006
|
10.1103/vsz9-gwvs
| null |
hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The supersymmetric custodial triplet model (SCTM), which is a fully-super\-symmetric generalization of the Georgi-Machacek (GM) model, is constructed by extending the Higgs sector of the minimal supersymmetric standard model by three triplet chiral superfields with hypercharge $Y=0,\pm 1$, in order to maintain the holomorphy of the superpotential and satisfy the requirements of anomaly cancellation. The global $SU(2)_L\otimes SU(2)_R$ symmetry has to be respected for the superpotential and soft supersymmetry breaking sector, where the former will only be broken by the Yukawa couplings and the latter one will be broken spontaneously to the custodial $SU(2)_V$ symmetry after electroweak symmetry breaking similar as the GM model. The ensuing complicated spectrum not only gives rich collider phenomenology but also provides more available Weakly Interacting Massive Particle (WIMP) Dark Matter (DM) candidates. In this paper, we explore the viability of WIMP solutions for DM in the SCTM, by considering the increasingly stringent constraints from direct detection DM experiments. Our numerical simulations show that a significant portion of the SCTM parameter space remains viable despite these constraints and will be fully tested in future experiments.
|
[
{
"version": "v1",
"created": "Tue, 04 Mar 2025 14:16:11 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 10:39:46 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Shang",
"LiangLiang",
""
],
[
"Wang",
"Yuanping",
""
],
[
"Du",
"Xiaokang",
""
],
[
"Yang",
"Bingfang",
""
],
[
"Moretti",
"Stefano",
""
]
] |
2503.03222
|
Zhumei Wang
|
Zhumei Wang, Zechen Hu, Ruoxi Guo, Huaijin Pi, Ziyong Feng, Sida Peng, Xiaowei Zhou, Mingtao Pei, Siyuan Huang
|
Mocap-2-to-3: Multi-view Lifting for Monocular Motion Recovery with 2D Pretraining
|
Project page: https://wangzhumei.github.io/mocap-2-to-3/
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recovering absolute human motion from monocular inputs is challenging due to two main issues. First, existing methods depend on 3D training data collected from limited environments, constraining out-of-distribution generalization. The second issue is the difficulty of estimating metric-scale poses from monocular input. To address these challenges, we introduce Mocap-2-to-3, a novel framework that performs multi-view lifting from monocular input by leveraging 2D data pre-training, enabling the reconstruction of metrically accurate 3D motions with absolute positions. To leverage abundant 2D data, we decompose complex 3D motion into multi-view syntheses. We first pretrain a single-view diffusion model on extensive 2D datasets, then fine-tune a multi-view model using public 3D data to enable view-consistent motion generation from monocular input, allowing the model to acquire action priors and diversity through 2D data. Furthermore, to recover absolute poses, we propose a novel human motion representation that decouples the learning of local pose and global movements, while encoding geometric priors of the ground to accelerate convergence. This enables progressive recovery of motion in absolute space during inference. Experimental results on in-the-wild benchmarks demonstrate that our method surpasses state-of-the-art approaches in both camera-space motion realism and world-grounded human positioning, while exhibiting superior generalization capability. Our code will be made publicly available.
|
[
{
"version": "v1",
"created": "Wed, 05 Mar 2025 06:32:49 GMT"
},
{
"version": "v2",
"created": "Thu, 06 Mar 2025 14:32:49 GMT"
},
{
"version": "v3",
"created": "Sun, 06 Apr 2025 13:54:00 GMT"
},
{
"version": "v4",
"created": "Mon, 28 Jul 2025 06:36:59 GMT"
},
{
"version": "v5",
"created": "Thu, 31 Jul 2025 11:03:35 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Zhumei",
""
],
[
"Hu",
"Zechen",
""
],
[
"Guo",
"Ruoxi",
""
],
[
"Pi",
"Huaijin",
""
],
[
"Feng",
"Ziyong",
""
],
[
"Peng",
"Sida",
""
],
[
"Zhou",
"Xiaowei",
""
],
[
"Pei",
"Mingtao",
""
],
[
"Huang",
"Siyuan",
""
]
] |
2503.03449
|
Giammarco Caroleo
|
Giammarco Caroleo, Alessandro Albini, Daniele De Martini, Timothy D. Barfoot, Perla Maiolino
|
Tiny LiDARs for Manipulator Self-Awareness: Sensor Characterization and Initial Localization Experiments
|
7 pages, 6 figures, 3 tables, IEEE/RSJ International Conference on Intelligent Robots and Systems 2025 accepted paper
| null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
For several tasks, ranging from manipulation to inspection, it is beneficial for robots to localize a target object in their surroundings. In this paper, we propose an approach that utilizes coarse point clouds obtained from miniaturized VL53L5CX Time-of-Flight (ToF) sensors (tiny LiDARs) to localize a target object in the robot's workspace. We first conduct an experimental campaign to calibrate the dependency of sensor readings on relative range and orientation to targets. We then propose a probabilistic sensor model, which we validate in an object pose estimation task using a Particle Filter (PF). The results show that the proposed sensor model improves the performance of the localization of the target object with respect to two baselines: one that assumes measurements are free from uncertainty and one in which the confidence is provided by the sensor datasheet.
|
[
{
"version": "v1",
"created": "Wed, 05 Mar 2025 12:26:48 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:25:23 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Caroleo",
"Giammarco",
""
],
[
"Albini",
"Alessandro",
""
],
[
"De Martini",
"Daniele",
""
],
[
"Barfoot",
"Timothy D.",
""
],
[
"Maiolino",
"Perla",
""
]
] |
2503.03596
|
Ariel A. P\'erez
|
Rodrigo Lecaros, Ariel A. P\'erez and Manuel F. Prado
|
Carleman estimate for semi-discrete stochastic parabolic operators in arbitrary dimension and applications to controllability
| null | null | null | null |
math.OC
|
http://creativecommons.org/licenses/by/4.0/
|
This paper considers a semi-discrete forward stochastic parabolic operator with homogeneous Dirichlet conditions in arbitrary dimensions. We show the lack of null controllability for a spatial semi-discretization of a null-controllable stochastic parabolic system from any initial datum. However, by proving a new Carleman estimate for its semi-discrete backward stochastic adjoint system, we achieve a relaxed observability inequality, which is applied to derivative $\phi$-null controllability by duality arguments.
|
[
{
"version": "v1",
"created": "Wed, 05 Mar 2025 15:31:56 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:48:21 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Lecaros",
"Rodrigo",
""
],
[
"Pérez",
"Ariel A.",
""
],
[
"Prado",
"Manuel F.",
""
]
] |
2503.04165
|
Bodong Zhang
|
Bodong Zhang, Hamid Manoochehri, Xiwen Li, Beatrice S. Knudsen, Tolga Tasdizen
|
WeakSupCon: Weakly Supervised Contrastive Learning for Encoder Pre-training
|
Medical Image Computing and Computer Assisted Intervention (MICCAI) 2025 workshop on Efficient Medical AI
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Weakly supervised multiple instance learning (MIL) is a challenging task given that only bag-level labels are provided, while each bag typically contains multiple instances. This topic has been extensively studied in histopathological image analysis, where labels are usually available only at the whole slide image (WSI) level, while each WSI could be divided into thousands of small image patches for training. The dominant MIL approaches focus on feature aggregation and take fixed patch features as inputs. However, weakly supervised feature representation learning in MIL settings is always neglected. Those features used to be generated by self-supervised learning methods that do not utilize weak labels, or by foundation encoders pre-trained on other large datasets. In this paper, we propose a novel weakly supervised feature representation learning method called Weakly Supervised Contrastive Learning (WeakSupCon) that utilizes bag-level labels. In our method, we employ multi-task learning and define distinct contrastive losses for samples with different bag labels. Our experiments demonstrate that the features generated using WeakSupCon with limited computing resources significantly enhance MIL classification performance compared to self-supervised approaches across three datasets. Our WeakSupCon code is available at github.com/BzhangURU/Paper_WeakSupCon
|
[
{
"version": "v1",
"created": "Thu, 06 Mar 2025 07:25:43 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 23:37:48 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhang",
"Bodong",
""
],
[
"Manoochehri",
"Hamid",
""
],
[
"Li",
"Xiwen",
""
],
[
"Knudsen",
"Beatrice S.",
""
],
[
"Tasdizen",
"Tolga",
""
]
] |
2503.04351
|
Kim Kiehn
|
Kim Kiehn, Albin Ahlb\"ack, Kathl\'en Kohn
|
PLMP -- Point-Line Minimal Problems for Projective SfM
| null | null | null | null |
cs.CV math.AG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We completely classify all minimal problems for Structure-from-Motion (SfM) where arrangements of points and lines are fully observed by multiple uncalibrated pinhole cameras. We find 291 minimal problems, 73 of which have unique solutions and can thus be solved linearly. Two of the linear problems allow an arbitrary number of views, while all other minimal problems have at most 9 cameras. All minimal problems have at most 7 points and at most 12 lines. We compute the number of solutions of each minimal problem, as this gives a measurement of the problem's intrinsic difficulty, and find that these number are relatively low (e.g., when comparing with minimal problems for calibrated cameras). Finally, by exploring stabilizer subgroups of subarrangements, we develop a geometric and systematic way to 1) factorize minimal problems into smaller problems, 2) identify minimal problems in underconstrained problems, and 3) formally prove non-minimality.
|
[
{
"version": "v1",
"created": "Thu, 06 Mar 2025 11:49:43 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 11:39:45 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kiehn",
"Kim",
""
],
[
"Ahlbäck",
"Albin",
""
],
[
"Kohn",
"Kathlén",
""
]
] |
2503.05147
|
Tomoya Naito
|
Tomoya Naito, Yuto Hijikata, Juzo Zenihiro, Gianluca Col\`o, and Hiroyuki Sagawa
|
Mirror-skin thickness: a possible observable sensitive to the charge symmetry breaking energy density functional
|
14 pages, 13 figures, 2 tables
|
The European Physical Journal A 61, 177 (2025)
|
10.1140/epja/s10050-025-01646-x
|
RIKEN-iTHEMS-Report-25
|
nucl-th nucl-ex
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We propose a new observable, named the mirror-skin thickness, in order to extract the strength of the charge symmetry breaking (CSB) term in an energy density functional (EDF). The mirror-skin thickness of $ N = 20 $ isotones and $ Z = 20 $ isotopes is studied by using Hartree-Fock-Bogoliubov (HFB) calculations with various Skyrme EDFs and adding CSB and charge independence breaking (CIB) terms. It is shown that the mirror-skin thickness is sensitive only to the CSB EDF, but hardly depends on either the isospin symmetric part of the nuclear EDF or the CIB term. Therefore, this observable can be used to extract the magnitude of the CSB term in the EDF quantitatively, either from experimental data or ab initio calculations. We have studied the accuracy in the mirror-skin thickness that is needed to extract sensible information. Our study may also help to understand the inconsistency between the strength of the phenomenological CSB and that extracted from ab initio calculations [Naito et al. Nuovo. Cim. C 47, 52 (2024)]. Among possible mirror pairs for experimental study, we propose the mirror-skin thickness between $ {}^{42} \mathrm{Ca} $ and $ {}^{42} \mathrm{Ti} $, which could be accessed in future experiments in RIBF and/or FRIB.
|
[
{
"version": "v1",
"created": "Fri, 07 Mar 2025 05:08:35 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:10:12 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Naito",
"Tomoya",
""
],
[
"Hijikata",
"Yuto",
""
],
[
"Zenihiro",
"Juzo",
""
],
[
"Colò",
"Gianluca",
""
],
[
"Sagawa",
"Hiroyuki",
""
]
] |
2503.05727
|
Heajun An
|
Heajun An, Marcos Silva, Qi Zhang, Arav Singh, Minqian Liu, Xinyi Zhang, Sarvech Qadir, Sang Won Lee, Lifu Huang, Pamela J. Wisniewski and Jin-Hee Cho
|
Toward Integrated Solutions: A Systematic Interdisciplinary Review of Cybergrooming Research
| null | null | null | null |
cs.CY cs.CR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Cybergrooming exploits minors through online trust-building, yet research remains fragmented, limiting holistic prevention. Social sciences focus on behavioral insights, while computational methods emphasize detection, but their integration remains insufficient. This review systematically synthesizes both fields using the PRISMA framework to enhance clarity, reproducibility, and cross-disciplinary collaboration. Findings show that qualitative methods offer deep insights but are resource-intensive, machine learning models depend on data quality, and standard metrics struggle with imbalance and cultural nuances. By bridging these gaps, this review advances interdisciplinary cybergrooming research, guiding future efforts toward more effective prevention and detection strategies.
|
[
{
"version": "v1",
"created": "Tue, 18 Feb 2025 03:08:29 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:33:16 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"An",
"Heajun",
""
],
[
"Silva",
"Marcos",
""
],
[
"Zhang",
"Qi",
""
],
[
"Singh",
"Arav",
""
],
[
"Liu",
"Minqian",
""
],
[
"Zhang",
"Xinyi",
""
],
[
"Qadir",
"Sarvech",
""
],
[
"Lee",
"Sang Won",
""
],
[
"Huang",
"Lifu",
""
],
[
"Wisniewski",
"Pamela J.",
""
],
[
"Cho",
"Jin-Hee",
""
]
] |
2503.05911
|
Carson Sobolewski
|
Carson Sobolewski, Zhenjiang Mao, Kshitij Maruti Vejre, Ivan Ruchkin
|
Generalizable Image Repair for Robust Visual Control
|
8 pages, 4 figures, 2 tables, 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2025)
| null | null | null |
cs.RO cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Vision-based control relies on accurate perception to achieve robustness. However, image distribution changes caused by sensor noise, adverse weather, and dynamic lighting can degrade perception, leading to suboptimal control decisions. Existing approaches, including domain adaptation and adversarial training, improve robustness but struggle to generalize to unseen corruptions while introducing computational overhead. To address this challenge, we propose a real-time image repair module that restores corrupted images before they are used by the controller. Our method leverages generative adversarial models, specifically CycleGAN and pix2pix, for image repair. CycleGAN enables unpaired image-to-image translation to adapt to novel corruptions, while pix2pix exploits paired image data when available to improve the quality. To ensure alignment with control performance, we introduce a control-focused loss function that prioritizes perceptual consistency in repaired images. We evaluated our method in a simulated autonomous racing environment with various visual corruptions. The results show that our approach significantly improves performance compared to baselines, mitigating distribution shift and enhancing controller reliability.
|
[
{
"version": "v1",
"created": "Fri, 07 Mar 2025 20:16:40 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 16:39:27 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Sobolewski",
"Carson",
""
],
[
"Mao",
"Zhenjiang",
""
],
[
"Vejre",
"Kshitij Maruti",
""
],
[
"Ruchkin",
"Ivan",
""
]
] |
2503.06314
|
Heike Pfau
|
Ming-Hua Chang, Steffen Backes, Donghui Lu, Nicolas Gauthier, Makoto Hashimoto, Guan-Yu Chen, Hai-Hu Wen, Sung-Kwan Mo, Zhi-Xun Shen, Roser Valenti, Heike Pfau
|
Observation of Two Cascading Screening Processes in an Iron-based Superconductor
| null |
Communications Materials 6, 157 (2025)
|
10.1038/s43246-025-00881-5
|
RIKEN-iTHEMS-Report-25
|
cond-mat.str-el
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Understanding how renormalized quasiparticles emerge in strongly correlated electron materials provides a challenge for both experiment and theory. It has been predicted that distinctive spin and orbital screening mechanisms drive this process in multiorbital materials with strong Coulomb and Hund's interactions. Here, we provide the experimental evidence of both mechanisms from angle-resolved photoemission spectroscopy on RbFe$_2$As$_2$. We observe that the emergence of low-energy Fe 3$d_{xy}$ quasiparticles below 90K is tied to spin screening. A second process changes the spectral weight at high energies up to room temperature. Supported by theoretical calculations we attribute it to orbital screening of Fe 3d atomic excitations. These two cascading screening processes drive the temperature evolution from a bad metal to a correlated Fermi liquid.
|
[
{
"version": "v1",
"created": "Sat, 08 Mar 2025 19:17:51 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 17:37:50 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Chang",
"Ming-Hua",
""
],
[
"Backes",
"Steffen",
""
],
[
"Lu",
"Donghui",
""
],
[
"Gauthier",
"Nicolas",
""
],
[
"Hashimoto",
"Makoto",
""
],
[
"Chen",
"Guan-Yu",
""
],
[
"Wen",
"Hai-Hu",
""
],
[
"Mo",
"Sung-Kwan",
""
],
[
"Shen",
"Zhi-Xun",
""
],
[
"Valenti",
"Roser",
""
],
[
"Pfau",
"Heike",
""
]
] |
2503.06891
|
Xupeng Xie
|
Xupeng Xie, Ruoyu Geng, Jun Ma, Boyu Zhou
|
AKF-LIO: LiDAR-Inertial Odometry with Gaussian Map by Adaptive Kalman Filter
|
Submitted to IROS 2025 Conference, https://github.com/xpxie/AKF-LIO.git
| null | null | null |
cs.RO
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Existing LiDAR-Inertial Odometry (LIO) systems typically use sensor-specific or environment-dependent measurement covariances during state estimation, leading to laborious parameter tuning and suboptimal performance in challenging conditions (e.g., sensor degeneracy and noisy observations). Therefore, we propose an Adaptive Kalman Filter (AKF) framework that dynamically estimates time-varying noise covariances of LiDAR and Inertial Measurement Unit (IMU) measurements, enabling context-aware confidence weighting between sensors. During LiDAR degeneracy, the system prioritizes IMU data while suppressing contributions from unreliable inputs like moving objects or noisy point clouds. Furthermore, a compact Gaussian-based map representation is introduced to model environmental planarity and spatial noise. A correlated registration strategy ensures accurate plane normal estimation via pseudo-merge, even in unstructured environments like forests. Extensive experiments validate the robustness of the proposed system across diverse environments, including dynamic scenes and geometrically degraded scenarios. Our method achieves reliable localization results across all MARS-LVIG sequences and ranks 8th on the KITTI Odometry Benchmark. The code will be released at https://github.com/xpxie/AKF-LIO.git.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 03:37:07 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:36:19 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Xie",
"Xupeng",
""
],
[
"Geng",
"Ruoyu",
""
],
[
"Ma",
"Jun",
""
],
[
"Zhou",
"Boyu",
""
]
] |
2503.06981
|
Manjun Cui
|
Manjun Cui, Zhichao Zhang and Wei Yao
|
Graph Chirp Signal and Graph Fractional Vertex-Frequency Energy Distribution
| null | null | null | null |
eess.SP
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Graph signal processing (GSP) has emerged as a powerful framework for analyzing data on irregular domains. In recent years, many classical techniques in signal processing (SP) have been successfully extended to GSP. Among them, chirp signals play a crucial role in various SP applications. However, graph chirp signals have not been formally defined despite their importance. Here, we define graph chirp signals and establish a comprehensive theoretical framework for their analysis. We propose the graph fractional vertex--frequency energy distribution (GFED), which provides a powerful tool for processing and analyzing graph chirp signals. We introduce the general fractional graph distribution (GFGD), a generalized vertex--frequency distribution, and the reduced interference GFED, which can suppress cross-term interference and enhance signal clarity. Furthermore, we propose a novel method for detecting graph signals through GFED domain filtering, facilitating robust detection and analysis of graph chirp signals in noisy environments. Moreover, this method can be applied to real-world data for denoising more effective than some state-of-the-arts, further demonstrating its practical significance.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 06:56:49 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 13:34:26 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cui",
"Manjun",
""
],
[
"Zhang",
"Zhichao",
""
],
[
"Yao",
"Wei",
""
]
] |
2503.06989
|
Wenzhuo Xu
|
Wenzhuo Xu, Zhipeng Wei, Xiongtao Sun, Zonghao Ying, Deyue Zhang, Dongdong Yang, Xiangzheng Zhang, Quanchen Zou
|
Probabilistic Modeling of Jailbreak on Multimodal LLMs: From Quantification to Application
| null | null | null | null |
cs.CR cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Recently, Multimodal Large Language Models (MLLMs) have demonstrated their superior ability in understanding multimodal content. However, they remain vulnerable to jailbreak attacks, which exploit weaknesses in their safety alignment to generate harmful responses. Previous studies categorize jailbreaks as successful or failed based on whether responses contain malicious content. However, given the stochastic nature of MLLM responses, this binary classification of an input's ability to jailbreak MLLMs is inappropriate. Derived from this viewpoint, we introduce jailbreak probability to quantify the jailbreak potential of an input, which represents the likelihood that MLLMs generated a malicious response when prompted with this input. We approximate this probability through multiple queries to MLLMs. After modeling the relationship between input hidden states and their corresponding jailbreak probability using Jailbreak Probability Prediction Network (JPPN), we use continuous jailbreak probability for optimization. Specifically, we propose Jailbreak-Probability-based Attack (JPA) that optimizes adversarial perturbations on input image to maximize jailbreak probability, and further enhance it as Multimodal JPA (MJPA) by including monotonic text rephrasing. To counteract attacks, we also propose Jailbreak-Probability-based Finetuning (JPF), which minimizes jailbreak probability through MLLM parameter updates. Extensive experiments show that (1) (M)JPA yields significant improvements when attacking a wide range of models under both white and black box settings. (2) JPF vastly reduces jailbreaks by at most over 60\%. Both of the above results demonstrate the significance of introducing jailbreak probability to make nuanced distinctions among input jailbreak abilities.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 07:10:38 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 10:26:35 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Xu",
"Wenzhuo",
""
],
[
"Wei",
"Zhipeng",
""
],
[
"Sun",
"Xiongtao",
""
],
[
"Ying",
"Zonghao",
""
],
[
"Zhang",
"Deyue",
""
],
[
"Yang",
"Dongdong",
""
],
[
"Zhang",
"Xiangzheng",
""
],
[
"Zou",
"Quanchen",
""
]
] |
2503.07047
|
Yongle Zhang
|
Yongle Zhang, Yimin Liu, Yan Huang, Qiang Wu
|
Recovering Partially Corrupted Objects via Sketch-Guided Bidirectional Feature Interaction
|
13 pages. This work has been submitted to the IEEE for possible publication
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Text-guided diffusion models have achieved remarkable success in object inpainting by providing high-level semantic guidance through text prompts. However, they often lack precise pixel-level spatial control, especially in scenarios involving partially corrupted objects where critical uncorrupted cues remain. To overcome this limitation, sketch-guided methods have been introduced, using either indirect gradient modulation or direct sketch injection to improve structural control. Yet, existing approaches typically establish a one-way mapping from the sketch to the masked regions only, neglecting the contextual information from unmasked object areas. This leads to a disconnection between the sketch and the uncorrupted content, thereby causing sketch-guided inconsistency and structural mismatch. To tackle this challenge, we propose a sketch-guided bidirectional feature interaction framework built upon a pretrained Stable Diffusion model. Our bidirectional interaction features two complementary directions, context-to-sketch and sketch-to-inpainting, that enable fine-grained spatial control for partially corrupted object inpainting. In the context-to-sketch direction, multi-scale latents from uncorrupted object regions are propagated to the sketch branch to generate a visual mask that adapts the sketch features to the visible context and denoising progress. In the sketch-to-inpainting direction, a sketch-conditional affine transformation modulates the influence of sketch guidance based on the learned visual mask, ensuring consistency with uncorrupted object content. This interaction is applied at multiple scales within the encoder of the diffusion U-Net, enabling the model to restore object structures with enhanced spatial fidelity. Extensive experiments on two newly constructed benchmark datasets demonstrate that our approach outperforms state-of-the-art methods.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 08:34:31 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 06:44:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhang",
"Yongle",
""
],
[
"Liu",
"Yimin",
""
],
[
"Huang",
"Yan",
""
],
[
"Wu",
"Qiang",
""
]
] |
2503.07174
|
Patrick Pilch
|
Patrick Pilch, Laur Peedu, Urmas Nagel, Toomas R\~o\~om, Changqing Zhu, Yurii Skourski, Xianghan Xu, Robert J. Cava, and Zhe Wang
|
Quantum spin dynamics of the honeycomb magnet K$_2$Co$_2$TeO$_6$ in high magnetic fields
| null | null |
10.1103/dnpq-lnrm
| null |
cond-mat.str-el
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
We present terahertz spectroscopic measurements of quantum spin dynamics in the honeycomb magnet K$_2$Co$_2$TeO$_6$ as a function of temperature, polarization and in an external magnetic field applied in the honeycomb plane. Magnetic excitations are resolved below the magnetic ordering temperature of $T_\text{N}$ = 12 K. In the applied magnetic field, we reveal characteristic field dependence not only for the magnetic excitations observed at zero field, but also a rich set of modes emerging in finite fields. The observed magnetic excitations exhibit clear dependence on the terahertz polarization, and characteristic features at field-induced phase transitions consistent with our high-field magnetization data. We cannot evidently resolve a continuumlike feature, even when the long-range magnetic order is presumably suppressed in the strong magnetic field, indicating that a Kitaev-type interaction, if existing, is subleading in this compound.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 10:52:18 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Pilch",
"Patrick",
""
],
[
"Peedu",
"Laur",
""
],
[
"Nagel",
"Urmas",
""
],
[
"Rõõm",
"Toomas",
""
],
[
"Zhu",
"Changqing",
""
],
[
"Skourski",
"Yurii",
""
],
[
"Xu",
"Xianghan",
""
],
[
"Cava",
"Robert J.",
""
],
[
"Wang",
"Zhe",
""
]
] |
2503.07212
|
Alessandro Alocco
|
Alessandro Alocco, Andrea Celotto, Luca Fasolo, Bernardo Galvano, Emanuele Palumbo, Luca Oberto, Luca Callegaro, Felice Francesco Tafuri, Patrizia Livreri, Emanuele Enrico
|
Towards Quantifying Two-Mode Correlation Linewidths in Quantum Circuits
| null |
2025 International Workshop on Integrated Nonlinear Microwave and Millimetre-Wave Circuits (INMMIC)
|
10.1109/INMMIC64198.2025.10975393
| null |
quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
This paper aims to quantify the linewidth of two-mode correlations in Traveling Wave Parametric Amplifiers (TWPAs). Artifacts induced by data acquisition and processing, such as windowing effects and acquisition time, are examined to understand their influence on the linewidth estimation of these correlations. The findings underscore the significance of acquisition parameters in optimizing two-mode correlation measurements, enhancing device characterization for quantum applications.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 11:51:40 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Alocco",
"Alessandro",
""
],
[
"Celotto",
"Andrea",
""
],
[
"Fasolo",
"Luca",
""
],
[
"Galvano",
"Bernardo",
""
],
[
"Palumbo",
"Emanuele",
""
],
[
"Oberto",
"Luca",
""
],
[
"Callegaro",
"Luca",
""
],
[
"Tafuri",
"Felice Francesco",
""
],
[
"Livreri",
"Patrizia",
""
],
[
"Enrico",
"Emanuele",
""
]
] |
2503.07452
|
Stefan Ivi\'c
|
Karlo Jakac, Luka Lan\v{c}a, Ante Sikirica, Stefan Ivi\'c
|
Efficient data-driven flow modeling for accurate passive scalar advection in submesoscale domains
|
Initial version of the manuscript
| null |
10.1016/j.apor.2025.104699
| null |
physics.flu-dyn math.OC
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Knowing the sea surface velocity field is essential for various applications, such as search and rescue operations and oil spill monitoring, where understanding the movement of objects or substances is critical. However, obtaining an accurate approximation of these advection processes is challenging, even with modern measuring equipment, such as high-frequency radar or advanced simulations based on oceanic flow models. Therefore this paper presents a data-driven framework to approximate sea surface velocity from spatially distributed observations, thus enabling efficient probability advection modeling across submesoscale domains. The system approximates transient flows by leveraging quasi-steady flow assumptions. To overcome the limitations of point measurements in capturing domain-wide circulation, the method employs a fusion of two simplified 2D flow models to approximate submesoscale dynamics, enabling complete velocity field reconstruction from scattered data. To ensure reliable flow dynamics, the approach iteratively adjusts boundary conditions in numerical simulations to align the simulated flow with observations. Experimental validation in Kvarner Bay using GPS-tracked drifters confirmed the system's ability to replace computationally intensive transient simulations by approximating flow fields based on model simplifications. The results demonstrate its efficiency across domains, making it a practical tool for real-world submesoscale applications requiring swift passive scalar advection.
|
[
{
"version": "v1",
"created": "Mon, 10 Mar 2025 15:31:05 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Jakac",
"Karlo",
""
],
[
"Lanča",
"Luka",
""
],
[
"Sikirica",
"Ante",
""
],
[
"Ivić",
"Stefan",
""
]
] |
2503.08827
|
Donghee Lee
|
Donghee Lee, Hye-Sung Lee, Jaeok Yi
|
Synaptic Field Theory for Neural Networks
|
6 pages, 3 figures. Version accepted for publication in PRD
| null | null | null |
hep-th cond-mat.dis-nn hep-ph
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Theoretical understanding of deep learning remains elusive despite its empirical success. In this study, we propose a novel "synaptic field theory" that describes the training dynamics of synaptic weights and biases in the continuum limit. Unlike previous approaches, our framework treats synaptic weights and biases as fields and interprets their indices as spatial coordinates, with the training data acting as external sources. This perspective offers new insights into the fundamental mechanisms of deep learning and suggests a pathway for leveraging well-established field-theoretic techniques to study neural network training.
|
[
{
"version": "v1",
"created": "Tue, 11 Mar 2025 19:00:09 GMT"
},
{
"version": "v2",
"created": "Thu, 20 Mar 2025 14:39:41 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 04:53:55 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Lee",
"Donghee",
""
],
[
"Lee",
"Hye-Sung",
""
],
[
"Yi",
"Jaeok",
""
]
] |
2503.09215
|
Jian Zhu
|
Jian Zhu, Zhengyu Jia, Tian Gao, Jiaxin Deng, Shidi Li, Lang Zhang, Fu Liu, Peng Jia, Xianpeng Lang
|
Other Vehicle Trajectories Are Also Needed: A Driving World Model Unifies Ego-Other Vehicle Trajectories in Video Latent Space
|
8 pages, 7 figures
| null | null | null |
cs.CV cs.AI
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Advanced end-to-end autonomous driving systems predict other vehicles' motions and plan ego vehicle's trajectory. The world model that can foresee the outcome of the trajectory has been used to evaluate the autonomous driving system. However, existing world models predominantly emphasize the trajectory of the ego vehicle and leave other vehicles uncontrollable. This limitation hinders their ability to realistically simulate the interaction between the ego vehicle and the driving scenario. In this paper, we propose a driving World Model named EOT-WM, unifying Ego-Other vehicle Trajectories in videos for driving simulation. Specifically, it remains a challenge to match multiple trajectories in the BEV space with each vehicle in the video to control the video generation. We first project ego-other vehicle trajectories in the BEV space into the image coordinate for vehicle-trajectory match via pixel positions. Then, trajectory videos are encoded by the Spatial-Temporal Variational Auto Encoder to align with driving video latents spatially and temporally in the unified visual space. A trajectory-injected diffusion Transformer is further designed to denoise the noisy video latents for video generation with the guidance of ego-other vehicle trajectories. In addition, we propose a metric based on control latent similarity to evaluate the controllability of trajectories. Extensive experiments are conducted on the nuScenes dataset, and the proposed model outperforms the state-of-the-art method by 30% in FID and 55% in FVD. The model can also predict unseen driving scenes with self-produced trajectories.
|
[
{
"version": "v1",
"created": "Wed, 12 Mar 2025 10:02:18 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 08:07:46 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 10:11:56 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhu",
"Jian",
""
],
[
"Jia",
"Zhengyu",
""
],
[
"Gao",
"Tian",
""
],
[
"Deng",
"Jiaxin",
""
],
[
"Li",
"Shidi",
""
],
[
"Zhang",
"Lang",
""
],
[
"Liu",
"Fu",
""
],
[
"Jia",
"Peng",
""
],
[
"Lang",
"Xianpeng",
""
]
] |
2503.10410
|
Yuwen Du
|
Yuwen Du, Anning Hu, Zichen Chao, Yifan Lu, Junhao Ge, Genjia Liu, Weitao Wu, Lanjun Wang, Siheng Chen
|
RoCo-Sim: Enhancing Roadside Collaborative Perception through Foreground Simulation
| null | null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Roadside Collaborative Perception refers to a system where multiple roadside units collaborate to pool their perceptual data, assisting vehicles in enhancing their environmental awareness. Existing roadside perception methods concentrate on model design but overlook data issues like calibration errors, sparse information, and multi-view consistency, leading to poor performance on recent published datasets. To significantly enhance roadside collaborative perception and address critical data issues, we present the first simulation framework RoCo-Sim for road-side collaborative perception. RoCo-Sim is capable of generating diverse, multi-view consistent simulated roadside data through dynamic foreground editing and full-scene style transfer of a single image. RoCo-Sim consists of four components: (1) Camera Extrinsic Optimization ensures accurate 3D to 2D projection for roadside cameras; (2) A novel Multi-View Occlusion-Aware Sampler (MOAS) determines the placement of diverse digital assets within 3D space; (3) DepthSAM innovatively models foreground-background relationships from single-frame fixed-view images, ensuring multi-view consistency of foreground; and (4) Scalable Post-Processing Toolkit generates more realistic and enriched scenes through style transfer and other enhancements. RoCo-Sim significantly improves roadside 3D object detection, outperforming SOTA methods by 83.74 on Rcooper-Intersection and 83.12 on TUMTraf-V2X for AP70. RoCo-Sim fills a critical gap in roadside perception simulation. Code and pre-trained models will be released soon: https://github.com/duyuwen-duen/RoCo-Sim
|
[
{
"version": "v1",
"created": "Thu, 13 Mar 2025 14:33:42 GMT"
},
{
"version": "v2",
"created": "Fri, 25 Jul 2025 11:34:24 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 00:27:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Du",
"Yuwen",
""
],
[
"Hu",
"Anning",
""
],
[
"Chao",
"Zichen",
""
],
[
"Lu",
"Yifan",
""
],
[
"Ge",
"Junhao",
""
],
[
"Liu",
"Genjia",
""
],
[
"Wu",
"Weitao",
""
],
[
"Wang",
"Lanjun",
""
],
[
"Chen",
"Siheng",
""
]
] |
2503.11237
|
Rabimba Karanjai
|
Rabimba Karanjai, Sam Blackshear, Lei Xu, Weidong Shi
|
Collaboration is all you need: LLM Assisted Safe Code Translation
| null | null |
10.1145/3696630.3728521
| null |
cs.AI cs.CL cs.SE
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
This paper introduces UniTranslator, a visionary framework that re-imagines code translation as a collaborative endeavor among multiple, compact LLMs. By orchestrating the interaction of specialized agents, each focused on different aspects of the translation process and grounded in a deep understanding of programming concepts, UniTranslator achieves a level of accuracy and efficiency that rivals larger, monolithic models. Our preliminary evaluation demonstrates the potential of UniTranslator to overcome the limitations of existing approaches and unlock the power of smaller LLMs for complex code translation tasks. We explore the effectiveness of this dynamic multi-agent paradigm in handling diverse language pairs, including low-resource languages, and in mitigating common issues such as code artifacts and hallucinations through the use of Natural Language Inference (NLI) grounding and iterative feedback mechanisms
|
[
{
"version": "v1",
"created": "Fri, 14 Mar 2025 09:42:07 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Karanjai",
"Rabimba",
""
],
[
"Blackshear",
"Sam",
""
],
[
"Xu",
"Lei",
""
],
[
"Shi",
"Weidong",
""
]
] |
2503.11806
|
Christopher Xie
|
Christopher Xie, Armen Avetisyan, Henry Howard-Jenkins, Yawar Siddiqui, Julian Straub, Richard Newcombe, Vasileios Balntas, Jakob Engel
|
Human-in-the-Loop Local Corrections of 3D Scene Layouts via Infilling
|
Project page: https://www.projectaria.com/scenescript/
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
We present a novel human-in-the-loop approach to estimate 3D scene layout that uses human feedback from an egocentric standpoint. We study this approach through introduction of a novel local correction task, where users identify local errors and prompt a model to automatically correct them. Building on SceneScript, a state-of-the-art framework for 3D scene layout estimation that leverages structured language, we propose a solution that structures this problem as "infilling", a task studied in natural language processing. We train a multi-task version of SceneScript that maintains performance on global predictions while significantly improving its local correction ability. We integrate this into a human-in-the-loop system, enabling a user to iteratively refine scene layout estimates via a low-friction "one-click fix'' workflow. Our system enables the final refined layout to diverge from the training distribution, allowing for more accurate modelling of complex layouts.
|
[
{
"version": "v1",
"created": "Fri, 14 Mar 2025 18:45:19 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 22:34:42 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Xie",
"Christopher",
""
],
[
"Avetisyan",
"Armen",
""
],
[
"Howard-Jenkins",
"Henry",
""
],
[
"Siddiqui",
"Yawar",
""
],
[
"Straub",
"Julian",
""
],
[
"Newcombe",
"Richard",
""
],
[
"Balntas",
"Vasileios",
""
],
[
"Engel",
"Jakob",
""
]
] |
2503.12725
|
Soofiyan Atar
|
Soofiyan Atar, Xiao Liang, Calvin Joyce, Florian Richter, Wood Ricardo, Charles Goldberg, Preetham Suresh, Michael Yip
|
Humanoids in Hospitals: A Technical Study of Humanoid Robot Surrogates for Dexterous Medical Interventions
|
8 pages
| null | null | null |
cs.RO
|
http://creativecommons.org/licenses/by/4.0/
|
The increasing demand for healthcare workers, driven by aging populations and labor shortages, presents a significant challenge for hospitals. Humanoid robots have the potential to alleviate these pressures by leveraging their human-like dexterity and adaptability to assist in medical procedures. This work conducted an exploratory study on the feasibility of humanoid robots performing direct clinical tasks through teleoperation. A bimanual teleoperation system was developed for the Unitree G1 Humanoid Robot, integrating high-fidelity pose tracking, custom grasping configurations, and an impedance controller to safely and precisely manipulate medical tools. The system is evaluated in seven diverse medical procedures, including physical examinations, emergency interventions, and precision needle tasks. Our results demonstrate that humanoid robots can successfully replicate critical aspects of human medical assessments and interventions, with promising quantitative performance in ventilation and ultrasound-guided tasks. However, challenges remain, including limitations in force output for procedures requiring high strength and sensor sensitivity issues affecting clinical accuracy. This study highlights the potential and current limitations of humanoid robots in hospital settings and lays the groundwork for future research on robotic healthcare integration.
|
[
{
"version": "v1",
"created": "Mon, 17 Mar 2025 01:31:27 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 05:28:51 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Atar",
"Soofiyan",
""
],
[
"Liang",
"Xiao",
""
],
[
"Joyce",
"Calvin",
""
],
[
"Richter",
"Florian",
""
],
[
"Ricardo",
"Wood",
""
],
[
"Goldberg",
"Charles",
""
],
[
"Suresh",
"Preetham",
""
],
[
"Yip",
"Michael",
""
]
] |
2503.13544
|
Juhyeong Kim
|
Juhyeong Kim, Sungyoon Choi, Youngbin Lee, Yejin Kim, Yongmin Choi and Yongjae Lee
|
Decision by Supervised Learning with Deep Ensembles: A Practical Framework for Robust Portfolio Optimization
|
8 pages, 3 figures
| null | null | null |
cs.LG q-fin.CP q-fin.PM
|
http://creativecommons.org/licenses/by/4.0/
|
We propose Decision by Supervised Learning (DSL), a practical framework for robust portfolio optimization. DSL reframes portfolio construction as a supervised learning problem: models are trained to predict optimal portfolio weights, using cross-entropy loss and portfolios constructed by maximizing the Sharpe or Sortino ratio. To further enhance stability and reliability, DSL employs Deep Ensemble methods, substantially reducing variance in portfolio allocations. Through comprehensive backtesting across diverse market universes and neural architectures, shows superior performance compared to both traditional strategies and leading machine learning-based methods, including Prediction-Focused Learning and End-to-End Learning. We show that increasing the ensemble size leads to higher median returns and more stable risk-adjusted performance. The code is available at https://github.com/DSLwDE/DSLwDE.
|
[
{
"version": "v1",
"created": "Sun, 16 Mar 2025 10:57:45 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Mar 2025 23:42:02 GMT"
},
{
"version": "v3",
"created": "Fri, 25 Jul 2025 08:25:59 GMT"
},
{
"version": "v4",
"created": "Wed, 30 Jul 2025 23:25:16 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kim",
"Juhyeong",
""
],
[
"Choi",
"Sungyoon",
""
],
[
"Lee",
"Youngbin",
""
],
[
"Kim",
"Yejin",
""
],
[
"Choi",
"Yongmin",
""
],
[
"Lee",
"Yongjae",
""
]
] |
2503.13831
|
Josef Willsher
|
Josef Willsher, Johannes Knolle
|
Dynamics and stability of U(1) spin liquids beyond mean-field theory: Triangular-lattice $J_1$-$J_2$ Heisenberg model
|
9 pages; 7 figures
| null | null | null |
cond-mat.str-el
|
http://creativecommons.org/licenses/by-nc-nd/4.0/
|
Quantum spin liquids (QSLs) are long-range entangled phases of frustrated magnets exhibiting fractionalized spin excitations. In two dimensions, there is limited analytical understanding of their excitation spectra beyond parton mean-field theories, which fail to capture many features of the finite frequency dynamical response from recent experimental and numerical works. We use a self-consistent random phase approximation (RPA) for the $J_1$-$J_2$ Heiseneberg model on the triangular lattice to describe the strong spinon-spinon interactions of the U(1) Dirac QSL. We obtain quantitative results for the dynamical spin structure factor and phase diagram compatible with comprehensive numerical efforts. We extend the method to chiral QSLs, and discuss its broad range of applicability to other models and for describing inelastic neutron scattering experiments.
|
[
{
"version": "v1",
"created": "Tue, 18 Mar 2025 02:12:34 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:24:58 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Willsher",
"Josef",
""
],
[
"Knolle",
"Johannes",
""
]
] |
2503.14939
|
Tengjin Weng
|
Tengjin Weng, Jingyi Wang, Wenhao Jiang and Zhong Ming
|
VisNumBench: Evaluating Number Sense of Multimodal Large Language Models
|
accepted by ICCV 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Can Multimodal Large Language Models (MLLMs) develop an intuitive number sense similar to humans? Targeting this problem, we introduce Visual Number Benchmark (VisNumBench) to evaluate the number sense abilities of MLLMs across a wide range of visual numerical tasks. VisNumBench consists of about 1,900 multiple-choice question-answer pairs derived from both synthetic and real-world visual data, covering seven visual numerical attributes and four types of visual numerical estimation tasks. Our experiments on VisNumBench led to the following key findings: (i) The 17 MLLMs we tested, including open-source models such as Qwen2.5-VL and InternVL2.5, as well as proprietary models like GPT-4o and Gemini 2.0 Flash, perform significantly below human levels in number sense-related tasks. (ii) Multimodal mathematical models and multimodal chain-of-thought (CoT) models did not exhibit significant improvements in number sense abilities. (iii) Stronger MLLMs with larger parameter sizes and broader general abilities demonstrate modest gains in number sense abilities. We believe VisNumBench will serve as a valuable resource for the research community, encouraging further advancements in enhancing MLLMs' number sense abilities. Code and dataset are available at https://wwwtttjjj.github.io/VisNumBench/.
|
[
{
"version": "v1",
"created": "Wed, 19 Mar 2025 07:07:43 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 09:21:08 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Weng",
"Tengjin",
""
],
[
"Wang",
"Jingyi",
""
],
[
"Jiang",
"Wenhao",
""
],
[
"Ming",
"Zhong",
""
]
] |
2503.15232
|
Ashley Willis
|
Ashley P. Willis and Michael J. Burin
|
Longitudinal vortices in unsteady Taylor-Couette flow: solution to a 60-year-old mystery
|
13 pages, 7 figures. Improved text and figures, additional calculations
| null | null | null |
physics.flu-dyn
|
http://creativecommons.org/licenses/by/4.0/
|
Applying a sufficiently rapid start-stop to the outer cylinder of the Couette-Taylor system, structures approximately aligned with the axis were recorded in the classic work of Coles (1965). These short-lived rolls are oriented perpendicular to the classic Taylor-vortex rolls. In this work we report numerical observation of this instability, guided by a more recent experimental observation. The instability is shown to be related to an inflection in the azimuthal velocity profile, a finding consistent with the experimental observations of its emergence during the deceleration phase. Despite the transient nature of start-stop experiments, we show that the instability can be linked to that of the oscillating boundary layer problem of Stokes. There are several reasons why the instability may have remained elusive, both for experimental observation and intrinsic to the idealized system. We look in more detail at dependence on the radius ratio for the Taylor-Couette system and find that, in the case where the size of the rolls scales with the gap width, for radius ratios any lower than that used by Coles, R_i/R_o=0.874, the instability is quickly overrun by axisymmetric rolls of Gortler type.
|
[
{
"version": "v1",
"created": "Wed, 19 Mar 2025 14:10:32 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 21:42:30 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Willis",
"Ashley P.",
""
],
[
"Burin",
"Michael J.",
""
]
] |
2503.15299
|
Zorik Gekhman
|
Zorik Gekhman, Eyal Ben David, Hadas Orgad, Eran Ofek, Yonatan Belinkov, Idan Szpektor, Jonathan Herzig, Roi Reichart
|
Inside-Out: Hidden Factual Knowledge in LLMs
|
Accepted to COLM 2025
| null | null | null |
cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
This work presents a framework for assessing whether large language models (LLMs) encode more factual knowledge in their parameters than what they express in their outputs. While a few studies hint at this possibility, none has clearly defined or demonstrated this phenomenon. We first propose a formal definition of knowledge, quantifying it for a given question as the fraction of correct-incorrect answer pairs where the correct one is ranked higher. This gives rise to external and internal knowledge, depending on the information used to score individual answer candidates: either the model's observable token-level probabilities or its intermediate computations. Hidden knowledge arises when internal knowledge exceeds external knowledge. We then present a case study, applying this framework to three popular open-weights LLMs in a closed-book QA setup. Our results indicate that: (1) LLMs consistently encode more factual knowledge internally than what they express externally, with an average relative gap of 40%. (2) Surprisingly, some knowledge is so deeply hidden that a model can internally know an answer perfectly, yet fail to generate it even once, despite large-scale repeated sampling of 1,000 answers. This reveals fundamental limitations in the generation capabilities of LLMs, which (3) put a practical constraint on scaling test-time compute via repeated answer sampling in closed-book QA: significant performance improvements remain inaccessible because some answers are practically never sampled, yet if they were, we would be guaranteed to rank them first.
|
[
{
"version": "v1",
"created": "Wed, 19 Mar 2025 15:21:48 GMT"
},
{
"version": "v2",
"created": "Mon, 24 Mar 2025 01:31:35 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 14:38:33 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Gekhman",
"Zorik",
""
],
[
"David",
"Eyal Ben",
""
],
[
"Orgad",
"Hadas",
""
],
[
"Ofek",
"Eran",
""
],
[
"Belinkov",
"Yonatan",
""
],
[
"Szpektor",
"Idan",
""
],
[
"Herzig",
"Jonathan",
""
],
[
"Reichart",
"Roi",
""
]
] |
2503.15621
|
Sara Sarto
|
Federico Cocchi, Nicholas Moratelli, Davide Caffagni, Sara Sarto, Lorenzo Baraldi, Marcella Cornia, Rita Cucchiara
|
LLaVA-MORE: A Comparative Study of LLMs and Visual Backbones for Enhanced Visual Instruction Tuning
|
ICCV 2025 Workshop on What is Next in Multimodal Foundation Models
| null | null | null |
cs.CV cs.AI cs.CL cs.MM
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Recent progress in Multimodal Large Language Models (MLLMs) has highlighted the critical roles of both the visual backbone and the underlying language model. While prior work has primarily focused on scaling these components to billions of parameters, the trade-offs between model size, architecture, and performance remain underexplored. Additionally, inconsistencies in training data and evaluation protocols have hindered direct comparisons, making it difficult to derive optimal design choices. In this paper, we introduce LLaVA-MORE, a new family of MLLMs that integrates recent language models with diverse visual backbones. To ensure fair comparisons, we employ a unified training protocol applied consistently across all architectures. Our analysis systematically explores both small- and medium-scale LLMs -- including Phi-4, LLaMA-3.1, and Gemma-2 -- to evaluate multimodal reasoning, generation, and instruction following, while examining the relationship between model size and performance. Beyond evaluating the LLM impact on final results, we conduct a comprehensive study of various visual encoders, ranging from CLIP-based architectures to alternatives such as DINOv2, SigLIP, and SigLIP2. Additional experiments investigate the effects of increased image resolution and variations in pre-training datasets. Overall, our results provide insights into the design of more effective MLLMs, offering a reproducible evaluation framework that facilitates direct comparisons and can guide future model development. Our source code and trained models are publicly available at: https://github.com/aimagelab/LLaVA-MORE.
|
[
{
"version": "v1",
"created": "Wed, 19 Mar 2025 18:10:12 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 12:41:25 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Cocchi",
"Federico",
""
],
[
"Moratelli",
"Nicholas",
""
],
[
"Caffagni",
"Davide",
""
],
[
"Sarto",
"Sara",
""
],
[
"Baraldi",
"Lorenzo",
""
],
[
"Cornia",
"Marcella",
""
],
[
"Cucchiara",
"Rita",
""
]
] |
2503.15768
|
Alexandra DeLucia
|
Alexandra DeLucia and Mark Dredze
|
Can one size fit all?: Measuring Failure in Multi-Document Summarization Domain Transfer
| null | null | null | null |
cs.CL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
Abstractive multi-document summarization (MDS) is the task of automatically summarizing information in multiple documents, from news articles to conversations with multiple speakers. The training approaches for current MDS models can be grouped into four approaches: end-to-end with special pre-training ("direct"), chunk-then-summarize, extract-then-summarize, and inference with GPT-style models. In this work, we evaluate MDS models across training approaches, domains, and dimensions (reference similarity, quality, and factuality), to analyze how and why models trained on one domain can fail to summarize documents from another (News, Science, and Conversation) in the zero-shot domain transfer setting. We define domain-transfer "failure" as a decrease in factuality, higher deviation from the target, and a general decrease in summary quality. In addition to exploring domain transfer for MDS models, we examine potential issues with applying popular summarization metrics out-of-the-box.
|
[
{
"version": "v1",
"created": "Thu, 20 Mar 2025 00:57:38 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 23:19:16 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"DeLucia",
"Alexandra",
""
],
[
"Dredze",
"Mark",
""
]
] |
2503.15882
|
Zhiren He
|
Zhiren He, Prathap Kumar Jharapla, Nicolas Leconte, Jeil Jung, Guru Khalsa
|
Optical control of the moir\'e twist angle
| null | null | null | null |
cond-mat.mtrl-sci physics.optics
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this theoretical work, we propose an all-optical method for fast, precise manipulation of two-dimensional multilayers by transferring orbital angular momentum from phase-structured light (e.g. vortex beams) to a 2D material flake. We model the light-matter interaction, analyze the twist dynamics, and develop a phase diagram for optical twists by mapping the system onto an impulsively forced nonlinear pendulum. Our findings reveal rich dynamical responses spanning single- and multi-pulse twist angle control to (quasi)stable dynamical trajectories, and suggest a pathway for all-optical measurement of the twist potential energy. Aided by classical potential estimates for the interlayer energy and numerical simulation, we demonstrate the feasibility of this approach with hexagonal boron nitride bilayers and extend the results to dichalcogenides with first-principles calculations. These results can be generalized to other 2D multilayers, paving the way for scalable and customizable moir\'e electronics and photonics.
|
[
{
"version": "v1",
"created": "Thu, 20 Mar 2025 06:09:54 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 04:04:10 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"He",
"Zhiren",
""
],
[
"Jharapla",
"Prathap Kumar",
""
],
[
"Leconte",
"Nicolas",
""
],
[
"Jung",
"Jeil",
""
],
[
"Khalsa",
"Guru",
""
]
] |
2503.15897
|
Junho Kim
|
Junho Kim, Gwangtak Bae, Eun Sun Lee, Young Min Kim
|
Learning 3D Scene Analogies with Neural Contextual Scene Maps
|
Accepted to ICCV 2025
| null | null | null |
cs.CV cs.LG
|
http://creativecommons.org/licenses/by/4.0/
|
Understanding scene contexts is crucial for machines to perform tasks and adapt prior knowledge in unseen or noisy 3D environments. As data-driven learning is intractable to comprehensively encapsulate diverse ranges of layouts and open spaces, we propose teaching machines to identify relational commonalities in 3D spaces. Instead of focusing on point-wise or object-wise representations, we introduce 3D scene analogies, which are smooth maps between 3D scene regions that align spatial relationships. Unlike well-studied single instance-level maps, these scene-level maps smoothly link large scene regions, potentially enabling unique applications in trajectory transfer in AR/VR, long demonstration transfer for imitation learning, and context-aware object rearrangement. To find 3D scene analogies, we propose neural contextual scene maps, which extract descriptor fields summarizing semantic and geometric contexts, and holistically align them in a coarse-to-fine manner for map estimation. This approach reduces reliance on individual feature points, making it robust to input noise or shape variations. Experiments demonstrate the effectiveness of our approach in identifying scene analogies and transferring trajectories or object placements in diverse indoor scenes, indicating its potential for robotics and AR/VR applications. Project page including the code is available through this link: https://82magnolia.github.io/3d_scene_analogies/.
|
[
{
"version": "v1",
"created": "Thu, 20 Mar 2025 06:49:33 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 02:35:01 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Kim",
"Junho",
""
],
[
"Bae",
"Gwangtak",
""
],
[
"Lee",
"Eun Sun",
""
],
[
"Kim",
"Young Min",
""
]
] |
2503.16180
|
Rasmus Svejstrup Nielsen
|
Rasmus Svejstrup Nielsen, \'Angel Labordet \'Alvarez, Yvonne Tomm, Galina Gurieva, Andres Ortega-Guerrero, Joachim Breternitz, Lorenzo Bastonero, Nicola Marzari, Carlo Antonio Pignedoli, Susan Schorr, Mirjana Dimitrievska
|
BaZrS$_\text{3}$ Lights Up: The Interplay of Electrons, Photons, and Phonons in Strongly Luminescent Single Crystals
| null | null |
10.1002/adom.202500915
| null |
cond-mat.mtrl-sci
|
http://creativecommons.org/licenses/by/4.0/
|
Chalcogenide perovskites have emerged as a promising class of materials for the next generation of optoelectronic applications, with BaZrS$_\text{3}$ attracting significant attention due to its wide bandgap, earth-abundant composition, and thermal and chemical stability. However, previous studies have consistently reported weak and ambiguous photoluminescence (PL), regardless of synthesis method, raising questions about the intrinsic optoelectronic quality of this compound. In this work, we demonstrate strong, band-to-band-dominated PL at room temperature in high-quality BaZrS$_\text{3}$ single crystals, with a PL quantum yield of $\sim$0.005\%. Despite the narrow, single-component PL emission band, time-resolved PL measurements reveal a carrier lifetime of $1.0\pm0.2$ ns. To understand the origin of the strong PL and short carrier lifetime, we perform multiwavelength excitation and polarization-dependent Raman measurements, supported by first-principles lattice dynamics calculations. We identify all 23 theoretically predicted Raman-active modes and their symmetries, providing a comprehensive reference for future studies. Our results indicate that phonon-assisted carrier decay and nontrivial electron-phonon interactions contribute to the short carrier lifetimes, as evidenced by Raman spectroscopy and DFT calculations. Further studies on compositional variations or partial cation/anion substitutions could mitigate electron-phonon coupling and enhance carrier lifetimes. By establishing a detailed reference for the intrinsic vibrational and optoelectronic properties of BaZrS$_\text{3}$, this work paves the way for further advancements in chalcogenide perovskites for energy and optoelectronic technologies.
|
[
{
"version": "v1",
"created": "Thu, 20 Mar 2025 14:24:36 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 09:10:04 GMT"
},
{
"version": "v3",
"created": "Wed, 26 Mar 2025 08:12:59 GMT"
},
{
"version": "v4",
"created": "Mon, 07 Jul 2025 12:11:20 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Nielsen",
"Rasmus Svejstrup",
""
],
[
"Álvarez",
"Ángel Labordet",
""
],
[
"Tomm",
"Yvonne",
""
],
[
"Gurieva",
"Galina",
""
],
[
"Ortega-Guerrero",
"Andres",
""
],
[
"Breternitz",
"Joachim",
""
],
[
"Bastonero",
"Lorenzo",
""
],
[
"Marzari",
"Nicola",
""
],
[
"Pignedoli",
"Carlo Antonio",
""
],
[
"Schorr",
"Susan",
""
],
[
"Dimitrievska",
"Mirjana",
""
]
] |
2503.16819
|
Yuki Fuseya
|
Kazuki Koie, Rikako Yaguchi, Yuki Fuseya
|
Topological blocking at the Bi(111) surface due to surface relaxation
|
6 pages, 5 figures
|
Phys. Rev. B 111, L201303 (2025)
|
10.1103/PhysRevB.111.L201303
| null |
cond-mat.mtrl-sci cond-mat.mes-hall
|
http://creativecommons.org/licenses/by/4.0/
|
The topological characteristics of Bi and its alloys with Sb have fueled intense debate since the prediction of three-dimensional topological insulators. However, a definitive resolution has not been reached to date. Here, we provide theoretical evidence that surface relaxation conceals the underlying bulk topology of pure Bi. Using density functional theory calculations for thin Bi(111) films (up to 17 bilayers), we first demonstrate a substantial inter-bilayer expansion near the surface. Motivated by this finding, we extend our analysis to thick Bi(111) films (up to 250 bilayers) incorporating relaxation layers, within the framework of a relativistic empirical tight-binding model. Our results reveal that these relaxation layers topologically block the emergence of surface state and significantly suppress the one-particle spectrum of surface states, thereby obscuring the experimental identification of Bi's topological properties. This phenomenon, which we term "topological blocking", provides crucial insights into the long-standing difficulty of observing surface states of Bi(111) at the $\bar{M}$ point. Furthermore, it establishes a framework for understanding and predicting the topological behavior in systems where surface relaxation disrupts the bulk-edge correspondence.
|
[
{
"version": "v1",
"created": "Fri, 21 Mar 2025 03:20:09 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Koie",
"Kazuki",
""
],
[
"Yaguchi",
"Rikako",
""
],
[
"Fuseya",
"Yuki",
""
]
] |
2503.17526
|
S\'ebastien Quetin
|
S\'ebastien Quetin, Tapotosh Ghosh, Farhad Maleki
|
Beyond the Encoder: Joint Encoder-Decoder Contrastive Pre-Training Improves Dense Prediction
| null | null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by/4.0/
|
Contrastive learning methods in self-supervised settings have primarily focused on pre-training encoders, while decoders are typically introduced and trained separately for downstream dense prediction tasks. However, this conventional approach overlooks the potential benefits of jointly pre-training both encoder and decoder. In this paper, we propose DeCon, an efficient encoder-decoder self-supervised learning (SSL) framework that supports joint contrastive pre-training. We first extend existing SSL architectures to accommodate diverse decoders and their corresponding contrastive losses. Then, we introduce a weighted encoder-decoder contrastive loss with non-competing objectives to enable the joint pre-training of encoder-decoder architectures. By adapting an established contrastive SSL framework for dense prediction tasks, DeCon achieves new state-of-the-art results: on COCO object detection and instance segmentation when pre-trained on COCO dataset; across almost all dense downstream benchmark tasks when pre-trained on COCO+ and ImageNet-1K. Our results demonstrate that joint pre-training enhances the representation power of the encoder and improves performance in dense prediction tasks. This gain persists across heterogeneous decoder architectures, various encoder architectures, and in out-of-domain limited-data scenarios.
|
[
{
"version": "v1",
"created": "Fri, 21 Mar 2025 20:19:13 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 16:37:43 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Quetin",
"Sébastien",
""
],
[
"Ghosh",
"Tapotosh",
""
],
[
"Maleki",
"Farhad",
""
]
] |
2503.17564
|
Vishwesh Ramanathan
|
Vishwesh Ramanathan, Tony Xu, Pushpak Pati, Faruk Ahmed, Maged Goubran, Anne L. Martel
|
ModalTune: Fine-Tuning Slide-Level Foundation Models with Multi-Modal Information for Multi-task Learning in Digital Pathology
| null | null | null | null |
eess.IV cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Prediction tasks in digital pathology are challenging due to the massive size of whole-slide images (WSIs) and the weak nature of training signals. Advances in computing, data availability, and self-supervised learning (SSL) have paved the way for slide-level foundation models (SLFMs) that can improve prediction tasks in low-data regimes. However, current methods under-utilize shared information between tasks and modalities. To overcome this challenge, we propose ModalTune, a novel fine-tuning framework which introduces the Modal Adapter to integrate new modalities without modifying SLFM weights. Additionally, we use large-language models (LLMs) to encode labels as text, capturing semantic relationships across multiple tasks and cancer types in a single training recipe. ModalTune achieves state-of-the-art (SOTA) results against both uni-modal and multi-modal models across four cancer types, jointly improving survival and cancer subtype prediction while remaining competitive in pan-cancer settings. Additionally, we show ModalTune is generalizable to two out-of-distribution (OOD) datasets. To our knowledge, this is the first unified fine-tuning framework for multi-modal, multi-task, and pan-cancer modeling in digital pathology.
|
[
{
"version": "v1",
"created": "Fri, 21 Mar 2025 22:50:09 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 22:10:58 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ramanathan",
"Vishwesh",
""
],
[
"Xu",
"Tony",
""
],
[
"Pati",
"Pushpak",
""
],
[
"Ahmed",
"Faruk",
""
],
[
"Goubran",
"Maged",
""
],
[
"Martel",
"Anne L.",
""
]
] |
2503.17788
|
Gaoge Han
|
Gaoge Han, Yongkang Cheng, Zhe Chen, Shaoli Huang, Tongliang Liu
|
Learning to Align and Refine: A Foundation-to-Diffusion Framework for Occlusion-Robust Two-Hand Reconstruction
| null | null | null | null |
cs.CV cs.AI
|
http://creativecommons.org/licenses/by-nc-sa/4.0/
|
Two-hand reconstruction from monocular images faces persistent challenges due to complex and dynamic hand postures and occlusions, causing significant difficulty in achieving plausible interaction alignment. Existing approaches struggle with such alignment issues, often resulting in misalignment and penetration artifacts. To tackle this, we propose a dual-stage Foundation-to-Diffusion framework that precisely align 2D prior guidance from vision foundation models and diffusion-based generative 3D interaction refinement to achieve occlusion-robust two-hand reconstruction. First, we introduce a lightweight fusion alignment encoder that aligns fused multimodal 2D priors like key points, segmentation maps, and depth cues from vision foundation models during training. This provides robust structured guidance, further enabling efficient inference without heavy foundation model encoders at test time while maintaining high reconstruction accuracy. Second, we implement a two-hand diffusion model explicitly trained to convert interpenetrated 3D poses into plausible, penetration-free counterparts. Through collision gradient-guided denoising, the model rectifies artifacts while preserving natural spatial relationships between hands. Extensive evaluations demonstrate that our method achieves state-of-the-art performance on InterHand2.6M, HIC, and FreiHAND datasets, significantly advancing occlusion handling and interaction robustness. Our code will be publicly released.
|
[
{
"version": "v1",
"created": "Sat, 22 Mar 2025 14:42:27 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 17:55:56 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Han",
"Gaoge",
""
],
[
"Cheng",
"Yongkang",
""
],
[
"Chen",
"Zhe",
""
],
[
"Huang",
"Shaoli",
""
],
[
"Liu",
"Tongliang",
""
]
] |
2503.17856
|
Radu Beche
|
Radu Beche, Sergiu Nedevschi
|
ClaraVid: A Holistic Scene Reconstruction Benchmark From Aerial Perspective With Delentropy-Based Complexity Profiling
|
Accepted ICCV 2025
| null | null | null |
cs.CV
|
http://creativecommons.org/licenses/by-sa/4.0/
|
The development of aerial holistic scene understanding algorithms is hindered by the scarcity of comprehensive datasets that enable both semantic and geometric reconstruction. While synthetic datasets offer an alternative, existing options exhibit task-specific limitations, unrealistic scene compositions, and rendering artifacts that compromise real-world applicability. We introduce ClaraVid, a synthetic aerial dataset specifically designed to overcome these limitations. Comprising 16,917 high-resolution images captured at 4032x3024 from multiple viewpoints across diverse landscapes, ClaraVid provides dense depth maps, panoptic segmentation, sparse point clouds, and dynamic object masks, while mitigating common rendering artifacts. To further advance neural reconstruction, we introduce the Delentropic Scene Profile (DSP), a novel complexity metric derived from differential entropy analysis, designed to quantitatively assess scene difficulty and inform reconstruction tasks. Utilizing DSP, we systematically benchmark neural reconstruction methods, uncovering a consistent, measurable correlation between scene complexity and reconstruction accuracy. Empirical results indicate that higher delentropy strongly correlates with increased reconstruction errors, validating DSP as a reliable complexity prior. The data and code are available on the project page at https://rdbch.github.io/claravid/
|
[
{
"version": "v1",
"created": "Sat, 22 Mar 2025 20:26:20 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 16:56:06 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Beche",
"Radu",
""
],
[
"Nedevschi",
"Sergiu",
""
]
] |
2503.18601
|
Hyelin Choi
|
Hyelin Choi, Woocheol Choi
|
A Linear Convergence Result for the Jacobi-Proximal Alternating Direction Method of Multipliers
|
25 pages, 24 figures
| null | null | null |
math.OC
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we analyze the convergence rate of the Jacobi-Proximal Alternating Direction Method of Multipliers (ADMM) initially introduced by Deng et al. for the block-structured optimization problem with linear constraint. The algorithm is well-suited for parallel implementation and widely used for large-scale multi-block optimization problems. While the o(1/k) convergence of the Jacobi-Proximal ADMM for the case $N \geq 3$ has been well-established in the previous work, to the best of our knowledge, its linear convergence for $N \geq 3$ remains unproven. We establish the linear convergence of the algorithm when the cost functions are strongly convex and smooth. Numerical experiments are presented supporting the convergence result.
|
[
{
"version": "v1",
"created": "Mon, 24 Mar 2025 12:00:56 GMT"
},
{
"version": "v2",
"created": "Sun, 30 Mar 2025 08:38:01 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 09:14:01 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Choi",
"Hyelin",
""
],
[
"Choi",
"Woocheol",
""
]
] |
2503.18666
|
Christopher M. Poskitt
|
Haoyu Wang and Christopher M. Poskitt and Jun Sun
|
AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents
|
Accepted by the 48th IEEE/ACM International Conference on Software Engineering (ICSE 2026)
| null | null | null |
cs.AI cs.CL
|
http://creativecommons.org/licenses/by/4.0/
|
Agents built on LLMs are increasingly deployed across diverse domains, automating complex decision-making and task execution. However, their autonomy introduces safety risks, including security vulnerabilities, legal violations, and unintended harmful actions. Existing mitigation methods, such as model-based safeguards and early enforcement strategies, fall short in robustness, interpretability, and adaptability. To address these challenges, we propose AgentSpec, a lightweight domain-specific language for specifying and enforcing runtime constraints on LLM agents. With AgentSpec, users define structured rules that incorporate triggers, predicates, and enforcement mechanisms, ensuring agents operate within predefined safety boundaries. We implement AgentSpec across multiple domains, including code execution, embodied agents, and autonomous driving, demonstrating its adaptability and effectiveness. Our evaluation shows that AgentSpec successfully prevents unsafe executions in over 90% of code agent cases, eliminates all hazardous actions in embodied agent tasks, and enforces 100% compliance by autonomous vehicles (AVs). Despite its strong safety guarantees, AgentSpec remains computationally lightweight, with overheads in milliseconds. By combining interpretability, modularity, and efficiency, AgentSpec provides a practical and scalable solution for enforcing LLM agent safety across diverse applications. We also automate the generation of rules using LLMs and assess their effectiveness. Our evaluation shows that the rules generated by OpenAI o1 achieve a precision of 95.56% and recall of 70.96% for embodied agents, successfully identify 87.26% of the risky code, and prevent AVs from breaking laws in 5 out of 8 scenarios.
|
[
{
"version": "v1",
"created": "Mon, 24 Mar 2025 13:31:48 GMT"
},
{
"version": "v2",
"created": "Mon, 07 Apr 2025 10:57:45 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 04:00:48 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Wang",
"Haoyu",
""
],
[
"Poskitt",
"Christopher M.",
""
],
[
"Sun",
"Jun",
""
]
] |
2503.18711
|
Michelle Jou
|
Thomas Sugg, Kyle O'Brien, Lekh Poudel, Alex Dumouchelle, Michelle Jou, Marc Bosch, Deva Ramanan, Srinivasa Narasimhan, Shubham Tulsiani
|
Accenture-NVS1: A Novel View Synthesis Dataset
|
6 pages, 7 figures
| null | null | null |
cs.CV cs.LG
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
This paper introduces ACC-NVS1, a specialized dataset designed for research on Novel View Synthesis specifically for airborne and ground imagery. Data for ACC-NVS1 was collected in Austin, TX and Pittsburgh, PA in 2023 and 2024. The collection encompasses six diverse real-world scenes captured from both airborne and ground cameras, resulting in a total of 148,000 images. ACC-NVS1 addresses challenges such as varying altitudes and transient objects. This dataset is intended to supplement existing datasets, providing additional resources for comprehensive research, rather than serving as a benchmark.
|
[
{
"version": "v1",
"created": "Mon, 24 Mar 2025 14:24:08 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 20:46:33 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Sugg",
"Thomas",
""
],
[
"O'Brien",
"Kyle",
""
],
[
"Poudel",
"Lekh",
""
],
[
"Dumouchelle",
"Alex",
""
],
[
"Jou",
"Michelle",
""
],
[
"Bosch",
"Marc",
""
],
[
"Ramanan",
"Deva",
""
],
[
"Narasimhan",
"Srinivasa",
""
],
[
"Tulsiani",
"Shubham",
""
]
] |
2503.18946
|
Yang-Zhi Chou
|
Jihang Zhu, Yang-Zhi Chou, Yi Huang, Sankar Das Sarma
|
In-plane magnetic field-induced orbital FFLO superconductivity in twisted WSe$_2$ homobilayers
|
7+5 pages, 4+5 figures. Published version
|
Phys. Rev. B 112, L020507 (2025); Editors' suggestion
|
10.1103/bqns-1ld3
| null |
cond-mat.supr-con cond-mat.mes-hall cond-mat.mtrl-sci cond-mat.str-el
|
http://creativecommons.org/licenses/by/4.0/
|
We theoretically predict the in-plane magnetic field-induced orbital Fulde-Ferrell-Larkin-Ovchinnikov (FFLO) superconducting states in twisted WSe$_2$ homobilayers (tWSe$_2$), focusing on its dependence on layer polarization and Fermi surface geometry. For unpolarized layers, finite-momentum pairing emerges only at low temperatures and above a critical field $B_{c1,\parallel}$. When layer symmetry is broken, finite-momentum pairing is stabilized at any nonzero field, with a critical temperature higher than that of the zero-momentum state. Notably, we identify a phase transition, which separates two distinct FFLO phases, when there are two separate Fermi pockets residing in the two moir\'e mini-valleys associated with opposite layers. We further discuss the effects of twist angles and applied field directions. Our findings establish tWSe$_2$ as a promising platform for realizing and manipulating FFLO states via twist angle, displacement field, and filling factor.
|
[
{
"version": "v1",
"created": "Mon, 24 Mar 2025 17:59:55 GMT"
},
{
"version": "v2",
"created": "Thu, 31 Jul 2025 15:07:04 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Zhu",
"Jihang",
""
],
[
"Chou",
"Yang-Zhi",
""
],
[
"Huang",
"Yi",
""
],
[
"Sarma",
"Sankar Das",
""
]
] |
2503.19008
|
Jose Tarciso Junior
|
Ednaldo L. B. Junior, Jos\'e Tarciso S. S. Junior, Francisco S. N. Lobo, Manuel E. Rodrigues, Lu\'is F. Dias da Silva and Henrique A. Vieira
|
Black hole solutions in Cotton gravity coupled to nonlinear electrodynamics
|
V2: 16 pages, 17 figures; discussion and references added. Accepted for publication in Physics of the Dark Universe
|
Physics of the Dark Universe 49 (2025) 102028
|
10.1016/j.dark.2025.102028
| null |
gr-qc astro-ph.CO hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The framework of General Relativity (GR) has recently been expanded through the introduction of Cotton Gravity (CG), a theoretical extension proposed by J. Harada. This modified approach integrates the Cotton tensor into the gravitational field equations, naturally encompassing all conventional GR solutions while allowing the cosmological constant to emerge as an integration constant. In this study, we delve into the implications of CG when coupled with nonlinear electrodynamics (NLED), constructing and analyzing three distinct static, spherically symmetric configurations. Our investigation centers on the horizon structure, metric characteristics, and the underlying NLED Lagrangian density of each model. We also confront the theoretical predictions with observational data by comparing the calculated shadow radii of these solutions to constraints imposed by the Event Horizon Telescope's (EHT) measurements of Sgr A$^{*}$. The results reveal an extensive spectrum of spacetime geometries, ranging from multi-horizon structures to naked singularities. Furthermore, the agreement between the predicted shadow sizes and EHT observations reinforces the viability of these models in describing the astrophysical image of Sgr A$^{*}$ within certain parameter bounds.
|
[
{
"version": "v1",
"created": "Mon, 24 Mar 2025 17:43:50 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Jul 2025 19:33:50 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Junior",
"Ednaldo L. B.",
""
],
[
"Junior",
"José Tarciso S. S.",
""
],
[
"Lobo",
"Francisco S. N.",
""
],
[
"Rodrigues",
"Manuel E.",
""
],
[
"da Silva",
"Luís F. Dias",
""
],
[
"Vieira",
"Henrique A.",
""
]
] |
2503.19480
|
Shijie Ma
|
Shijie Ma, Yuying Ge, Teng Wang, Yuxin Guo, Yixiao Ge, Ying Shan
|
GenHancer: Imperfect Generative Models are Secretly Strong Vision-Centric Enhancers
|
ICCV 2025. Project released at: https://mashijie1028.github.io/GenHancer/
| null | null | null |
cs.CV
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
The synergy between generative and discriminative models receives growing attention. While discriminative Contrastive Language-Image Pre-Training (CLIP) excels in high-level semantics, it struggles with perceiving fine-grained visual details. Generally, to enhance representations, generative models take CLIP's visual features as conditions for reconstruction. However, the underlying principle remains underexplored. In this work, we empirically found that visually perfect generations are not always optimal for representation enhancement. The essence lies in effectively extracting fine-grained knowledge from generative models while mitigating irrelevant information. To explore critical factors, we delve into three aspects: (1) Conditioning mechanisms: We found that even a small number of local tokens can drastically reduce the difficulty of reconstruction, leading to collapsed training. We thus conclude that utilizing only global visual tokens as conditions is the most effective strategy. (2) Denoising configurations: We observed that end-to-end training introduces extraneous information. To address this, we propose a two-stage training strategy to prioritize learning useful visual knowledge. Additionally, we demonstrate that lightweight denoisers can yield remarkable improvements. (3) Generation paradigms: We explore both continuous and discrete denoisers with desirable outcomes, validating the versatility of our method. Through our in-depth explorations, we have finally arrived at an effective method, namely GenHancer, which consistently outperforms prior arts on the MMVP-VLM benchmark, e.g., 6.0% on OpenAICLIP. The enhanced CLIP can be further plugged into multimodal large language models for better vision-centric performance. All the models and codes are made publicly available.
|
[
{
"version": "v1",
"created": "Tue, 25 Mar 2025 09:15:34 GMT"
},
{
"version": "v2",
"created": "Fri, 30 May 2025 18:23:14 GMT"
},
{
"version": "v3",
"created": "Thu, 31 Jul 2025 13:18:52 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Ma",
"Shijie",
""
],
[
"Ge",
"Yuying",
""
],
[
"Wang",
"Teng",
""
],
[
"Guo",
"Yuxin",
""
],
[
"Ge",
"Yixiao",
""
],
[
"Shan",
"Ying",
""
]
] |
2503.19848
|
Alex Glynn
|
Alex Glynn
|
Guarding against artificial intelligence--hallucinated citations: the case for full-text reference deposit
|
3 pages
|
Glynn A. Guarding against artificial intelligence -- hallucinated citations: The case for full-text reference deposit. Eur Sci Ed. 2025;51:e153973
|
10.3897/ese.2025.e153973
| null |
cs.DL cs.AI
|
http://creativecommons.org/licenses/by/4.0/
|
The tendency of generative artificial intelligence (AI) systems to "hallucinate" false information is well-known; AI-generated citations to non-existent sources have made their way into the reference lists of peer-reviewed publications. Here, I propose a solution to this problem, taking inspiration from the Transparency and Openness Promotion (TOP) data sharing guidelines, the clash of generative AI with the American judiciary, and the precedent set by submissions of prior art to the United States Patent and Trademark Office. Journals should require authors to submit the full text of each cited source along with their manuscripts, thereby preventing authors from citing any material whose full text they cannot produce. This solution requires limited additional work on the part of authors or editors while effectively immunizing journals against hallucinated references.
|
[
{
"version": "v1",
"created": "Tue, 25 Mar 2025 17:12:38 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Glynn",
"Alex",
""
]
] |
2503.21034
|
Hidekazu Tanaka
|
Hidekazu Tanaka and Shuji Sasagawa
|
Properties of solutions by the Schwinger-Dyson equation at finite temperature and density : A four-fermion interaction model
|
30 pages, 25 figures
| null | null | null |
hep-ph hep-th
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
In this paper, we examine the properties of the solutions obtained by the Schwinger-Dyson equation (SDE).
As a simple example, we consider a two-dimensional model including four-fermion interaction. It is shown that when this model is solved by an iterative method using the SDE at finite density, multiple solutions depending on the initial input values are obtained.
We investigate the reasons for this situation by examining the convergence of the solutions obtained by iterative method. We also consider the solution of the SDE using alternative methods. Furthermore, we compare these results with analytical solutions at zero temperature and discuss the relation with the behavior of the effective potential. We extend these considerations to finite temperatures.
Based on these analyses, we consider ways to avoid spurious solutions that appear in the iterative method.
|
[
{
"version": "v1",
"created": "Wed, 26 Mar 2025 22:48:23 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 20:59:22 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Tanaka",
"Hidekazu",
""
],
[
"Sasagawa",
"Shuji",
""
]
] |
2503.21090
|
Joseph Carolan
|
Joseph Carolan, Andrew M. Childs, Matt Kovacs-Deak, Luke Schaeffer
|
Translation-Invariant Quantum Algorithms for Ordered Search are Optimal
|
31 pages, 1 figure. Generalized main result to bounded error algorithms
| null | null | null |
quant-ph
|
http://creativecommons.org/licenses/by/4.0/
|
Ordered search is the task of finding an item in an ordered list using comparison queries. The best exact classical algorithm for this fundamental problem uses $\lceil \log_{2}{n}\rceil$ queries for a list of length $n$. Quantum computers can achieve a constant-factor speedup, but the best possible coefficient of $\log_{2}{n}$ for exact quantum algorithms is only known to lie between $(\ln{2})/\pi \approx 0.221$ and $4/\log_{2}{605} \approx 0.433$. We consider a special class of translation-invariant algorithms with no workspace, introduced by Farhi, Goldstone, Gutmann, and Sipser, that has been used to find the best known upper bounds. First, we show that any bounded-error, $k$-query quantum algorithm for ordered search can be implemented by a $k$-query algorithm in this special class. Second, we use linear programming to show that the best exact $5$-query quantum algorithm can search a list of length $7265$, giving an ordered search algorithm that asymptotically uses $5 \log_{7265}{n} \approx 0.390 \log_{2}{n}$ quantum queries.
|
[
{
"version": "v1",
"created": "Thu, 27 Mar 2025 02:08:16 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Jul 2025 18:25:19 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Carolan",
"Joseph",
""
],
[
"Childs",
"Andrew M.",
""
],
[
"Kovacs-Deak",
"Matt",
""
],
[
"Schaeffer",
"Luke",
""
]
] |
2503.21813
|
Zhangcheng Qiang
|
Zhangcheng Qiang, Kerry Taylor, Weiqing Wang, Jing Jiang
|
OAEI-LLM-T: A TBox Benchmark Dataset for Understanding Large Language Model Hallucinations in Ontology Matching
|
14 pages, 4 figures, 4 tables, 2 prompt templates
| null | null | null |
cs.CL cs.AI cs.IR
|
http://arxiv.org/licenses/nonexclusive-distrib/1.0/
|
Hallucinations are often inevitable in downstream tasks using large language models (LLMs). To tackle the substantial challenge of addressing hallucinations for LLM-based ontology matching (OM) systems, we introduce a new benchmark dataset OAEI-LLM-T. The dataset evolves from seven TBox datasets in the Ontology Alignment Evaluation Initiative (OAEI), capturing hallucinations of ten different LLMs performing OM tasks. These OM-specific hallucinations are organised into two primary categories and six sub-categories. We showcase the usefulness of the dataset in constructing an LLM leaderboard for OM tasks and for fine-tuning LLMs used in OM tasks.
|
[
{
"version": "v1",
"created": "Tue, 25 Mar 2025 18:20:04 GMT"
},
{
"version": "v2",
"created": "Wed, 07 May 2025 15:02:02 GMT"
},
{
"version": "v3",
"created": "Wed, 14 May 2025 16:57:13 GMT"
}
] | 2025-08-01T00:00:00 |
[
[
"Qiang",
"Zhangcheng",
""
],
[
"Taylor",
"Kerry",
""
],
[
"Wang",
"Weiqing",
""
],
[
"Jiang",
"Jing",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.