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
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.05259 | Guillaume Pourcel | Guillaume Pourcel, Maxence Ernoult | Learning long range dependencies through time reversal symmetry breaking | 45 pages | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Deep State Space Models (SSMs) reignite physics-grounded compute paradigms, as RNNs could natively be embodied into dynamical systems. This calls for dedicated learning algorithms obeying to core physical principles, with efficient techniques to simulate these systems and guide their design. We propose Recurrent Hamiltonian Echo Learning (RHEL), an algorithm which provably computes loss gradients as finite differences of physical trajectories of non-dissipative, Hamiltonian systems. In ML terms, RHEL only requires three "forward passes" irrespective of model size, without explicit Jacobian computation, nor incurring any variance in the gradient estimation. Motivated by the physical realization of our algorithm, we first introduce RHEL in continuous time and demonstrate its formal equivalence with the continuous adjoint state method. To facilitate the simulation of Hamiltonian systems trained by RHEL, we propose a discrete-time version of RHEL which is equivalent to Backpropagation Through Time (BPTT) when applied to a class of recurrent modules which we call Hamiltonian Recurrent Units (HRUs). This setting allows us to demonstrate the scalability of RHEL by generalizing these results to hierarchies of HRUs, which we call Hamiltonian SSMs (HSSMs). We apply RHEL to train HSSMs with linear and nonlinear dynamics on a variety of time-series tasks ranging from mid-range to long-range classification and regression with sequence length reaching $\sim 50k$. We show that RHEL consistently matches the performance of BPTT across all models and tasks. This work opens new doors for the design of scalable, energy-efficient physical systems endowed with self-learning capabilities for sequence modelling. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:20:39 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Pourcel",
"Guillaume",
""
],
[
"Ernoult",
"Maxence",
""
]
] |
2506.05260 | Xiaodong Wang | Xiaodong Wang, Jinfa Huang, Li Yuan, Peixi Peng | LeanPO: Lean Preference Optimization for Likelihood Alignment in Video-LLMs | Code: https://github.com/Wang-Xiaodong1899/LeanPO | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Most Video Large Language Models (Video-LLMs) adopt preference alignment techniques, e.g., DPO~\citep{rafailov2024dpo}, to optimize the reward margin between a winning response ($y_w$) and a losing response ($y_l$). However, the likelihood displacement observed in DPO indicates that both $\log \pi_\theta (y_w\mid x)$ and $\log \pi_\theta (y_l\mid x) $ often decrease during training, inadvertently boosting the probabilities of non-target responses. In this paper, we systematically revisit this phenomenon from LLMs to Video-LLMs, showing that it intensifies when dealing with the redundant complexity of video content. To alleviate the impact of this phenomenon, we propose \emph{Lean Preference Optimization} (LeanPO), a reference-free approach that reformulates the implicit reward as the average likelihood of the response with respect to the policy model. A key component of LeanPO is the reward-trustworthiness correlated self-generated preference data pipeline, which carefully infuses relevant prior knowledge into the model while continuously refining the preference data via self-reflection. This allows the policy model to obtain high-quality paired data and accurately estimate the newly defined reward, thus mitigating the unintended drop. In addition, we introduce a dynamic label smoothing strategy that mitigates the impact of noise in responses from diverse video content, preventing the model from overfitting to spurious details. Extensive experiments demonstrate that LeanPO significantly enhances the performance of state-of-the-art Video-LLMs, consistently boosting baselines of varying capacities with minimal additional training overhead. Moreover, LeanPO offers a simple yet effective solution for aligning Video-LLM preferences with human trustworthiness, paving the way toward the reliable and efficient Video-LLMs. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:21:16 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wang",
"Xiaodong",
""
],
[
"Huang",
"Jinfa",
""
],
[
"Yuan",
"Li",
""
],
[
"Peng",
"Peixi",
""
]
] |
2506.05261 | Rick Danielson | Richard E. Danielson, Minghong Zhang, Jo\"el Chass\'e, Will Perrie | A seasonal to decadal calibration of 1990-2100 eastern Canadian freshwater discharge simulations by observations, data models, and neural networks | null | null | null | null | stat.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A configuration of the NCAR WRF-Hydro model was sought using well established data models to guide the initial hydrologic model setup, as well as a seasonal streamflow post-processing by neural networks. Discharge was simulated using an eastern Canadian river network at two-km resolution. The river network was taken from a digital elevation model that was made to conform to observed catchment boundaries. Perturbations of a subset of model parameters were examined with reference to streamflow from 25 gauged catchments during the 2019 warm season. A data model defines the similarity of modelled streamflow to observations, and improvements were found in about half the individual catchments. With reference to 183 gauged catchments (1990-2022), further improvements were obtained at monthly and annual scales by neural network post-processing that targets all catchments at once as well as individual catchments.
This seasonal calibration was applied to uncoupled WRF-Hydro simulations for the 1990-2100 warming period. Historic and future forcing were provided, respectively, by a European Centre for Medium-Range Weather Forecasting reanalysis (ERA5), and by a WRF atmospheric model downscaling of a set of Coupled Model Intercomparison Project (CMIP) models, where the latter were also seasonally calibrated. Eastern Canadian freshwater discharge peaks at about 10$^5$ m$^3$ s$^{-1}$, and as previous studies have shown, there is a trend toward increasing low flows during the cold season and an earlier peak discharge in spring. By design, neural networks yield more precise estimates by compensating for different hydrologic process representations. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:21:23 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Danielson",
"Richard E.",
""
],
[
"Zhang",
"Minghong",
""
],
[
"Chassé",
"Joël",
""
],
[
"Perrie",
"Will",
""
]
] |
2506.05262 | Amir Shee | Amir Shee, Ritwik Bandyopadhyay and Haicen Yue | Tuning Shear Rheology through Active Dopants | 7 pages, 5 figures | null | null | null | cond-mat.soft cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | We numerically investigate the shear rheology of mixtures of active and passive Brownian particles, with varying fractions of active components. We find that even a small fraction of active dopants triggers fluidization with comparable efficiency to fully active systems. A combined parameter, active energy, given by dopant fraction multiplied by propulsion speed squared controls the shear rheology and glass transition of the active-passive mixtures. These results together provide a quantitative strategy for fine-tuning the mechanical properties of a soft material with small amounts of active dopants. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:21:24 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Shee",
"Amir",
""
],
[
"Bandyopadhyay",
"Ritwik",
""
],
[
"Yue",
"Haicen",
""
]
] |
2506.05263 | Juan Tapia Dr. | Juan E. Tapia and Christoph Busch | Can Foundation Models Generalise the Presentation Attack Detection Capabilities on ID Cards? | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Nowadays, one of the main challenges in presentation attack detection (PAD) on ID cards is obtaining generalisation capabilities for a diversity of countries that are issuing ID cards. Most PAD systems are trained on one, two, or three ID documents because of privacy protection concerns. As a result, they do not obtain competitive results for commercial purposes when tested in an unknown new ID card country. In this scenario, Foundation Models (FM) trained on huge datasets can help to improve generalisation capabilities. This work intends to improve and benchmark the capabilities of FM and how to use them to adapt the generalisation on PAD of ID Documents. Different test protocols were used, considering zero-shot and fine-tuning and two different ID card datasets. One private dataset based on Chilean IDs and one open-set based on three ID countries: Finland, Spain, and Slovakia. Our findings indicate that bona fide images are the key to generalisation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:24:11 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Tapia",
"Juan E.",
""
],
[
"Busch",
"Christoph",
""
]
] |
2506.05264 | Pradeep Kumar Mohanty | Abir Bhowmick, Sayantan Mitra and P. K. Mohanty | Geometric and Nonequilibrium Criticality in Run-and-Tumble Particles with Competing Motility and Attraction | 14 pages, 12 figures (separate supplement.pdf attached) | null | null | null | cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Self-propulsion in run-and-tumble particles (RTPs) generates effective attractive interactions that can drive motility-induced phase separation (MIPS), a phenomenon absent in their passive counterparts. In this work, we show that at high motility, introducing explicit attractive interactions among RTPs can suppress MIPS, leading to a homogeneous phase, and subsequently induce a re-emergence of phase separation at stronger attraction --thus realizing a reentrant phase transition. We characterize this transition by examining the percolation properties of dense clusters, which serve as geometric signatures of phase separation. Along the resulting critical line, we observe continuously varying critical exponents, while some of the associated scaling functions remain invariant and coincide with those of equilibrium lattice gas models undergoing interacting percolation, which is Ising-percolation universality. These findings reveal that the phase separation transition in interacting RTPs exhibits Ising-like super universality, bridging nonequilibrium active matter with classical critical behavior. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:24:17 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Bhowmick",
"Abir",
""
],
[
"Mitra",
"Sayantan",
""
],
[
"Mohanty",
"P. K.",
""
]
] |
2506.05265 | Mohammed Almutairi | Mohammed Almutairi | Teaming in the AI Era: AI-Augmented Frameworks for Forming, Simulating, and Optimizing Human Teams | 5 pages, UMAP 25, June 16_19, 2025, New York City, NY, USA | ACM International Conference on User Modeling, Adaptation and Personalization 2025 | 10.1145/3699682.3727574 | null | cs.HC cs.AI cs.MA | http://creativecommons.org/licenses/by/4.0/ | Effective teamwork is essential across diverse domains. During the team formation stage, a key challenge is forming teams that effectively balance user preferences with task objectives to enhance overall team satisfaction. In the team performing stage, maintaining cohesion and engagement is critical for sustaining high team performance. However, existing computational tools and algorithms for team optimization often rely on static data inputs, narrow algorithmic objectives, or solutions tailored for specific contexts, failing to account for the dynamic interplay of team members personalities, evolving goals, and changing individual preferences. Therefore, teams may encounter member dissatisfaction, as purely algorithmic assignments can reduce members commitment to team goals or experience suboptimal engagement due to the absence of timely, personalized guidance to help members adjust their behaviors and interactions as team dynamics evolve. Ultimately, these challenges can lead to reduced overall team performance. My Ph.D. dissertation aims to develop AI-augmented team optimization frameworks and practical systems that enhance team satisfaction, engagement, and performance. First, I propose a team formation framework that leverages a multi-armed bandit algorithm to iteratively refine team composition based on user preferences, ensuring alignment between individual needs and collective team goals to enhance team satisfaction. Second, I introduce tAIfa (Team AI Feedback Assistant), an AI-powered system that utilizes large language models (LLMs) to deliver immediate, personalized feedback to both teams and individual members, enhancing cohesion and engagement. Finally, I present PuppeteerLLM, an LLM-based simulation framework that simulates multi-agent teams to model complex team dynamics within realistic environments, incorporating task-driven collaboration and long-term coordination. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:24:37 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Almutairi",
"Mohammed",
""
]
] |
2506.05266 | Benjamin Doyon | Benjamin Doyon | Nonlinear projection for ballistic correlation functions: a formula in terms of minimal connected covers | 38 pages, 1 figure | null | null | null | cond-mat.stat-mech math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many-body systems, the dynamics is governed, at large scales of space and time, by the hydrodynamic principle of projection onto the conserved densities admitted by the model. This is formalised as local relaxation of fluctuations in the ballistic macroscopic fluctuation theory, a nonlinear Boltzmann-Gibbs principle. We use it to derive a projection formula, expressing n-point connected correlation functions (cumulants) of generic observables at different space-time points, in terms of those of conserved densities. This applies in every d >= 1 spatial dimensions and under the ballistic scaling of space and time, both in and out of equilibrium. It generalises the well-known linear-response principle for 2-point functions. For higher-point functions, one needs to account for nonlinear fluctuations of conserved densities. The result is a nonlinear projection, expressed as a sum over certain products of lower-order correlation functions of conserved densities with equilibrium multivariances as coefficients. Using Malyshev's formula, the sum is combinatorially organised via certain covers of the set of space-time points, which we call "minimal connected covers". We use this in order to get general, explicit formulas for two- and three-point functions in stationary states, expressed in terms of thermodynamic and Euler-scale data. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:26:03 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Doyon",
"Benjamin",
""
]
] |
2506.05267 | Van C. Nguyen | Nicol\'as Andruskiewitsch, David Jaklitsch, Van C. Nguyen, Amrei Oswald, Julia Plavnik, Anne V. Shepler, and Xingting Wang | On the finite generation of the cohomology of bosonizations | 33 pages, comments are welcome | null | null | null | math.QA math.KT math.RA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We use deformation sequences of (Hopf) algebras, extending the results of Negron and Pevtsova, to show that bosonizations of some suitable braided Hopf algebras by some suitable finite-dimensional Hopf algebras have finitely generated cohomology. In fact, our results are shown in more generality for smash products. As applications, we prove the bosonizations of some Nichols algebras (such as Nichols algebras of diagonal type, the restricted Jordan plane, Nichols algebras of direct sums of Jordan blocks plus points labeled with 1), by some suitable finite-dimensional Hopf algebras, have finitely generated cohomology, recovering some known results as well as providing new examples. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:26:39 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Andruskiewitsch",
"Nicolás",
""
],
[
"Jaklitsch",
"David",
""
],
[
"Nguyen",
"Van C.",
""
],
[
"Oswald",
"Amrei",
""
],
[
"Plavnik",
"Julia",
""
],
[
"Shepler",
"Anne V.",
""
],
[
"Wang",
"Xingting",
""
]
] |
2506.05268 | Selena Ling | Selena Ling, Abhishek Madan, Nicholas Sharp, Alec Jacobson | Uniform Sampling of Surfaces by Casting Rays | 15 pages, 17 figures, Symposium on Geometry Processing 2025 | null | null | null | cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Randomly sampling points on surfaces is an essential operation in geometry processing. This sampling is computationally straightforward on explicit meshes, but it is much more difficult on other shape representations, such as widely-used implicit surfaces. This work studies a simple and general scheme for sampling points on a surface, which is derived from a connection to the intersections of random rays with the surface. Concretely, given a subroutine to cast a ray against a surface and find all intersections, we can use that subroutine to uniformly sample white noise points on the surface. This approach is particularly effective in the context of implicit signed distance functions, where sphere marching allows us to efficiently cast rays and sample points, without needing to extract an intermediate mesh. We analyze the basic method to show that it guarantees uniformity, and find experimentally that it is significantly more efficient than alternative strategies on a variety of representations. Furthermore, we show extensions to blue noise sampling and stratified sampling, and applications to deform neural implicit surfaces as well as moment estimation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:26:48 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ling",
"Selena",
""
],
[
"Madan",
"Abhishek",
""
],
[
"Sharp",
"Nicholas",
""
],
[
"Jacobson",
"Alec",
""
]
] |
2506.05269 | Nicolas Mousset Mr | Nicolas Mousset and Nina H. Amini | State Space Decomposition of Quantum Dynamical Semigroups | This will be published in IEEE qCCL 2025 | null | null | null | quant-ph math-ph math.MP math.OC math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The mean evolution of an open quantum system in continuous time is described by a time continuous semigroup of quantum channels (completely positive and trace-preserving linear maps). Baumgartner and Narnhofer presented a general decomposition of the underlying Hilbert space into a sum of invariant subspaces, also called enclosures. We propose a new reading of this result, inspired by the work of Carbone and Pautrat. In addition, we apply this decomposition to a class of open quantum random walks and to quantum trajectories, where we study its uniqueness. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:28:22 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Mousset",
"Nicolas",
""
],
[
"Amini",
"Nina H.",
""
]
] |
2506.05270 | Nicola Picenni | Massimo Gobbino, Nicola Picenni | Symmetry breaking for local minimizers of a free discontinuity problem | 27 pages, 2 figures | null | null | null | math.AP math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study a functional defined on the class of piecewise constant functions, combining a jump penalization, which discourages discontinuities, with a fidelity term that penalizes deviations from a given linear function, called the forcing term.
In one dimension, it is not difficult to see that local minimizers form staircases that approximate the forcing term. Here we show that in two dimensions symmetry breaking occurs, leading to the emergence of exotic minimizers whose level sets are not simple stripes with boundaries orthogonal to the gradient of the forcing term.
The proof relies on the calibration method for free discontinuity problems. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:29:17 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Gobbino",
"Massimo",
""
],
[
"Picenni",
"Nicola",
""
]
] |
2506.05271 | Daniel Berg Thomsen | Daniel Berg Thomsen, Adrien Taylor, Aymeric Dieuleveut | Tight analyses of first-order methods with error feedback | null | null | null | null | cs.LG cs.DC math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Communication between agents often constitutes a major computational bottleneck in distributed learning. One of the most common mitigation strategies is to compress the information exchanged, thereby reducing communication overhead. To counteract the degradation in convergence associated with compressed communication, error feedback schemes -- most notably $\mathrm{EF}$ and $\mathrm{EF}^{21}$ -- were introduced. In this work, we provide a tight analysis of both of these methods. Specifically, we find the Lyapunov function that yields the best possible convergence rate for each method -- with matching lower bounds. This principled approach yields sharp performance guarantees and enables a rigorous, apples-to-apples comparison between $\mathrm{EF}$, $\mathrm{EF}^{21}$, and compressed gradient descent. Our analysis is carried out in a simplified yet representative setting, which allows for clean theoretical insights and fair comparison of the underlying mechanisms. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:30:18 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Thomsen",
"Daniel Berg",
""
],
[
"Taylor",
"Adrien",
""
],
[
"Dieuleveut",
"Aymeric",
""
]
] |
2506.05272 | Enric Ventura | Gemma Bastardas, Enric Ventura | Computing $H$-equations with 2-by-2 integral matrices | null | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | We study the transference through finite index extensions of the notion of equational coherence, as well as its effective counterpart. We deduce an explicit algorithm for solving the following algorithmic problem about size two integral invertible matrices: ''given $h_1,\ldots, h_r; g\in \operatorname{PSL}_2(\mathbb{Z})$, decide whether $g$ is algebraic over the subgroup $H=\langle h_1,\ldots ,h_r\rangle \leqslant \operatorname{PSL}_2(\mathbb{Z})$ (i.e., whether there exist a non-trivial $H$-equation $w(x)\in H*\langle x\rangle$ such that $w(g)=1$) and, in the affirmative case, compute finitely many such $H$-equations $w_1(x),\ldots ,w_s(x)\in H*\langle x\rangle$ further satisfying that any $w(x)\in H*\langle x\rangle$ with $w(g)=1$ is a product of conjugates of $w_1(x),\ldots ,w_s(x)$''. The same problem for square matrices of size 4 and bigger is unsolvable. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:30:49 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Bastardas",
"Gemma",
""
],
[
"Ventura",
"Enric",
""
]
] |
2506.05273 | Yerong Xu | Yerong Xu, Luigi C. Gallo, Kouichi Hagino, James N. Reeves, Francesco Tombesi, Misaki Mizumoto, Alfredo Luminari, Adam G. Gonzalez, Ehud Behar, Rozenn Boissay-Malaquin, Valentina Braito, Pierpaolo Condo, Chris Done, Aiko Miyamoto, Ryuki Mizukawa, Hirokazu Odaka, Riki Sato, Atsushi Tanimoto, Makoto Tashiro, Tahir Yaqoob, and Satoshi Yamada | Unraveling the structure of the stratified ultra-fast outflows in PDS 456 with XRISM | 22 pages, 18 figures, and 2 tables, accepted for publication in PASJ for the XRISM special issue | null | null | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Multiple clumpy wind components ($v_{out}\sim0.2-0.3c$) in the luminous quasar PDS 456 have recently been resolved by XRISM in the Fe-K band for the first time. In this paper, we investigate the structure of ultra-fast outflows (UFOs) using coordinated observations from XRISM, XMM-Newton, and NuSTAR, along with the self-consistently calculated photoionization model \texttt{PION}. Our results reveal a stratified ionization structure likely driven by the radiation field, characterized by a relation between wind velocity and ionization parameter $v_{out}\propto\xi^{(0.38\pm0.06)}$. To evaluate the impact of the screening effect, we tested all possible order permutations of six \texttt{PION} components. We find that highly ionized UFOs ($\log\xi>4.5$) are insensitive to their relative positions, whereas the soft X-ray UFO ($\log\xi\sim3$ and $v_{out}\sim0.27c$) and the lowest-ionized hard X-ray UFO ($\log\xi\sim4.1$ and $v_ {out}\sim0.23c$) are statistically favored -- based on the evidence from both the C-statistic and Bayesian analysis -- to occupy the middle and innermost layers, respectively. This suggests a possible trend where slower UFOs are launched from regions closer to the supermassive black hole (SMBH). The soft X-ray UFO is found to be thermally unstable, regardless of its relative position. However, its location remains unclear. Our sequence analysis and its similarity to hard X-ray UFOs suggest that they may be co-spatial, while variability constraints support its location within the broad-line region at sub-parsec scales. Simulations with the gate-valve opened XRISM show that high-resolution soft X-ray data can enhance the reliability of our results. Furthermore, simulations with the future X-ray mission NewAthena demonstrate its capability to resolve the absorber sequence and spatial distributions, enabling the determination of UFO structures and their roles in AGN feedback. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:30:58 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Xu",
"Yerong",
""
],
[
"Gallo",
"Luigi C.",
""
],
[
"Hagino",
"Kouichi",
""
],
[
"Reeves",
"James N.",
""
],
[
"Tombesi",
"Francesco",
""
],
[
"Mizumoto",
"Misaki",
""
],
[
"Luminari",
"Alfredo",
""
],
[
"Gonzalez",
"Adam G.",
""
],
[
"Behar",
"Ehud",
""
],
[
"Boissay-Malaquin",
"Rozenn",
""
],
[
"Braito",
"Valentina",
""
],
[
"Condo",
"Pierpaolo",
""
],
[
"Done",
"Chris",
""
],
[
"Miyamoto",
"Aiko",
""
],
[
"Mizukawa",
"Ryuki",
""
],
[
"Odaka",
"Hirokazu",
""
],
[
"Sato",
"Riki",
""
],
[
"Tanimoto",
"Atsushi",
""
],
[
"Tashiro",
"Makoto",
""
],
[
"Yaqoob",
"Tahir",
""
],
[
"Yamada",
"Satoshi",
""
]
] |
2506.05274 | Animesh Gupta | Animesh Gupta, Jay Parmar, Ishan Rajendrakumar Dave, Mubarak Shah | From Play to Replay: Composed Video Retrieval for Temporally Fine-Grained Videos | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Composed Video Retrieval (CoVR) retrieves a target video given a query video and a modification text describing the intended change. Existing CoVR benchmarks emphasize appearance shifts or coarse event changes and therefore do not test the ability to capture subtle, fast-paced temporal differences. We introduce TF-CoVR, the first large-scale benchmark dedicated to temporally fine-grained CoVR. TF-CoVR focuses on gymnastics and diving and provides 180K triplets drawn from FineGym and FineDiving. Previous CoVR benchmarks focusing on temporal aspect, link each query to a single target segment taken from the same video, limiting practical usefulness. In TF-CoVR, we instead construct each <query, modification> pair by prompting an LLM with the label differences between clips drawn from different videos; every pair is thus associated with multiple valid target videos (3.9 on average), reflecting real-world tasks such as sports-highlight generation. To model these temporal dynamics we propose TF-CoVR-Base, a concise two-stage training framework: (i) pre-train a video encoder on fine-grained action classification to obtain temporally discriminative embeddings; (ii) align the composed query with candidate videos using contrastive learning. We conduct the first comprehensive study of image, video, and general multimodal embedding (GME) models on temporally fine-grained composed retrieval in both zero-shot and fine-tuning regimes. On TF-CoVR, TF-CoVR-Base improves zero-shot mAP@50 from 5.92 (LanguageBind) to 7.51, and after fine-tuning raises the state-of-the-art from 19.83 to 25.82. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:31:17 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Gupta",
"Animesh",
""
],
[
"Parmar",
"Jay",
""
],
[
"Dave",
"Ishan Rajendrakumar",
""
],
[
"Shah",
"Mubarak",
""
]
] |
2506.05275 | Patrik \v{S}van\v{c}ara | Sreelekshmi C. Ajithkumar, Vitor S. Barroso, Patrik \v{S}van\v{c}ara, Anthony J. Kent, Silke Weinfurtner | Optical lever for broadband detection of fluid interface fluctuations | 14 pages, 13 figures | null | null | null | physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We exploit the optical lever principle to detect minute fluctuations of a liquid-air interface. Waves propagating on the interface deflect a specularly reflected laser beam, inducing angular deviations captured by a dual-element photodiode. We realise this principle in a compact set-up including a temperature-controlled fluid sample. Deflection angle fluctuations span five orders of magnitude in frequency, enabling detection of both low-frequency eigenmodes and high-frequency capillary waves driven by thermal effects. These results demonstrate the broad dynamical range and versatility of specular reflection spectroscopy as a minimally-invasive tool for probing interfacial dynamics in fluid and soft matter systems. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:31:49 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ajithkumar",
"Sreelekshmi C.",
""
],
[
"Barroso",
"Vitor S.",
""
],
[
"Švančara",
"Patrik",
""
],
[
"Kent",
"Anthony J.",
""
],
[
"Weinfurtner",
"Silke",
""
]
] |
2506.05276 | Hao Yu | Hao Yu, Chu Xin Cheng, Runlong Yu, Yuyang Ye, Shiwei Tong, Zhaofeng Liu, Defu Lian | How to Unlock Time Series Editing? Diffusion-Driven Approach with Multi-Grained Control | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent advances in time series generation have shown promise, yet controlling properties in generated sequences remains challenging. Time Series Editing (TSE) - making precise modifications while preserving temporal coherence - consider both point-level constraints and segment-level controls that current methods struggle to provide. We introduce the CocktailEdit framework to enable simultaneous, flexible control across different types of constraints. This framework combines two key mechanisms: a confidence-weighted anchor control for point-wise constraints and a classifier-based control for managing statistical properties such as sums and averages over segments. Our methods achieve precise local control during the denoising inference stage while maintaining temporal coherence and integrating seamlessly, with any conditionally trained diffusion-based time series models. Extensive experiments across diverse datasets and models demonstrate its effectiveness. Our work bridges the gap between pure generative modeling and real-world time series editing needs, offering a flexible solution for human-in-the-loop time series generation and editing. The code and demo are provided for validation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:32:00 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Yu",
"Hao",
""
],
[
"Cheng",
"Chu Xin",
""
],
[
"Yu",
"Runlong",
""
],
[
"Ye",
"Yuyang",
""
],
[
"Tong",
"Shiwei",
""
],
[
"Liu",
"Zhaofeng",
""
],
[
"Lian",
"Defu",
""
]
] |
2506.05277 | Arseny Shur | Arseny Shur | On Minimizers of Minimum Density | 30 pages, 4 figures | null | null | null | cs.DS cs.FL | http://creativecommons.org/licenses/by/4.0/ | Minimizers are sampling schemes with numerous applications in computational biology. Assuming a fixed alphabet of size $\sigma$, a minimizer is defined by two integers $k,w\ge2$ and a linear order $\rho$ on strings of length $k$ (also called $k$-mers). A string is processed by a sliding window algorithm that chooses, in each window of length $w+k-1$, its minimal $k$-mer with respect to $\rho$. A key characteristic of the minimizer is its density, which is the expected frequency of chosen $k$-mers among all $k$-mers in a random infinite $\sigma$-ary string. Minimizers of smaller density are preferred as they produce smaller samples with the same guarantee: each window is represented by a $k$-mer.
The problem of finding a minimizer of minimum density for given input parameters $(\sigma,k,w)$ has a huge search space of $(\sigma^k)!$ and is representable by an ILP of size $\tilde\Theta(\sigma^{k+w})$, which has worst-case solution time that is doubly-exponential in $(k+w)$ under standard complexity assumptions. We solve this problem in $w\cdot 2^{\sigma^k+O(k)}$ time and provide several additional tricks reducing the practical runtime and search space. As a by-product, we describe an algorithm computing the average density of a minimizer within the same time bound. Then we propose a novel method of studying minimizers via regular languages and show how to find, via the eigenvalue/eigenvector analysis over finite automata, minimizers with the minimal density in the asymptotic case $w\to\infty$. Implementing our algorithms, we compute the minimum density minimizers for $(\sigma,k)\in\{(2,2),(2,3),(2,4),(2,5),(4,2)\}$ and \textbf{all} $w\ge 2$. The obtained densities are compared against the average density and the theoretical lower bounds, including the new bound presented in this paper. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:33:00 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Shur",
"Arseny",
""
]
] |
2506.05278 | Nan Huo | Nan Huo, Jinyang Li, Bowen Qin, Ge Qu, Xiaolong Li, Xiaodong Li, Chenhao Ma, Reynold Cheng | Micro-Act: Mitigate Knowledge Conflict in Question Answering via Actionable Self-Reasoning | Accepted by ACL 2025 Main | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Retrieval-Augmented Generation (RAG) systems commonly suffer from Knowledge Conflicts, where retrieved external knowledge contradicts the inherent, parametric knowledge of large language models (LLMs). It adversely affects performance on downstream tasks such as question answering (QA). Existing approaches often attempt to mitigate conflicts by directly comparing two knowledge sources in a side-by-side manner, but this can overwhelm LLMs with extraneous or lengthy contexts, ultimately hindering their ability to identify and mitigate inconsistencies. To address this issue, we propose Micro-Act a framework with a hierarchical action space that automatically perceives context complexity and adaptively decomposes each knowledge source into a sequence of fine-grained comparisons. These comparisons are represented as actionable steps, enabling reasoning beyond the superficial context. Through extensive experiments on five benchmark datasets, Micro-Act consistently achieves significant increase in QA accuracy over state-of-the-art baselines across all 5 datasets and 3 conflict types, especially in temporal and semantic types where all baselines fail significantly. More importantly, Micro-Act exhibits robust performance on non-conflict questions simultaneously, highlighting its practical value in real-world RAG applications. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:33:02 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Huo",
"Nan",
""
],
[
"Li",
"Jinyang",
""
],
[
"Qin",
"Bowen",
""
],
[
"Qu",
"Ge",
""
],
[
"Li",
"Xiaolong",
""
],
[
"Li",
"Xiaodong",
""
],
[
"Ma",
"Chenhao",
""
],
[
"Cheng",
"Reynold",
""
]
] |
2506.05279 | Benjamin Doyon | Benjamin Doyon | Hydrodynamic noise in one dimension: projected Kubo formula and its vanishing in integrable models | 30 pages | null | null | null | cond-mat.stat-mech math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hydrodynamic noise is the Gaussian process that emerges at larges scales of space and time in many-body systems. It arises by the central limit theorem applied to local microcanonical averages, representing the degrees of freedom that have been forgotten when projecting coarse-grained observables onto conserved quantities. It comes with "bare" diffusion terms. In one dimension of space, nonlinearities of the hydrodynamic equation are relevant (from a renormalisation perspective), usually giving rise to hydrodynamic superdiffusion. But in linearly degenerate systems, where the relevant nonlinearity vanishes, the diffusive scaling stays intact. Nevertheless, anomalies remain. We show that in such systems, the noise covariance is determined in terms of a modification of the Kubo formula, where effects of ballistic long-range correlations have been subtracted. This is the projected Onsager matrix, in which so-called quadratic charges are projected out. We show that the Einstein relation holds, giving a projected bare diffusion, and that the remaining nonlinearities are tamed by a point-splitting regularisation. Putting these ingredients together, we obtain an exact and well-defined hydrodynamic fluctuation theory in the ballistic scaling of space-time, for the asymptotic expansion in the inverse variation scale, including the first subleading (diffusive-scale) corrections beyond large deviations. This is expressed as a stochastic PDE. We then obtain the anomalous hydrodynamic equation, which takes into account separately long-range correlations and bare diffusion. Using these result, in integrable systems, we show that hydrodynamic noise must be absent, as was conjectured recently. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:33:07 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Doyon",
"Benjamin",
""
]
] |
2506.05280 | Nan Wang | Nan Wang, Yuantao Chen, Lixing Xiao, Weiqing Xiao, Bohan Li, Zhaoxi Chen, Chongjie Ye, Shaocong Xu, Saining Zhang, Ziyang Yan, Pierre Merriaux, Lei Lei, Tianfan Xue and Hao Zhao | Unifying Appearance Codes and Bilateral Grids for Driving Scene Gaussian Splatting | Project page: https://bigcileng.github.io/bilateral-driving; Code: https://github.com/BigCiLeng/bilateral-driving | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural rendering techniques, including NeRF and Gaussian Splatting (GS), rely on photometric consistency to produce high-quality reconstructions. However, in real-world scenarios, it is challenging to guarantee perfect photometric consistency in acquired images. Appearance codes have been widely used to address this issue, but their modeling capability is limited, as a single code is applied to the entire image. Recently, the bilateral grid was introduced to perform pixel-wise color mapping, but it is difficult to optimize and constrain effectively. In this paper, we propose a novel multi-scale bilateral grid that unifies appearance codes and bilateral grids. We demonstrate that this approach significantly improves geometric accuracy in dynamic, decoupled autonomous driving scene reconstruction, outperforming both appearance codes and bilateral grids. This is crucial for autonomous driving, where accurate geometry is important for obstacle avoidance and control. Our method shows strong results across four datasets: Waymo, NuScenes, Argoverse, and PandaSet. We further demonstrate that the improvement in geometry is driven by the multi-scale bilateral grid, which effectively reduces floaters caused by photometric inconsistency. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:33:41 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wang",
"Nan",
""
],
[
"Chen",
"Yuantao",
""
],
[
"Xiao",
"Lixing",
""
],
[
"Xiao",
"Weiqing",
""
],
[
"Li",
"Bohan",
""
],
[
"Chen",
"Zhaoxi",
""
],
[
"Ye",
"Chongjie",
""
],
[
"Xu",
"Shaocong",
""
],
[
"Zhang",
"Saining",
""
],
[
"Yan",
"Ziyang",
""
],
[
"Merriaux",
"Pierre",
""
],
[
"Lei",
"Lei",
""
],
[
"Xue",
"Tianfan",
""
],
[
"Zhao",
"Hao",
""
]
] |
2506.05281 | Haifeng Sun | Haifeng Sun, Yu Xiong, Runze Wu, Xinyu Cai, Changjie Fan, Lan Zhang, Xiang-Yang Li | Fast-DataShapley: Neural Modeling for Training Data Valuation | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The value and copyright of training data are crucial in the artificial intelligence industry. Service platforms should protect data providers' legitimate rights and fairly reward them for their contributions. Shapley value, a potent tool for evaluating contributions, outperforms other methods in theory, but its computational overhead escalates exponentially with the number of data providers. Recent works based on Shapley values attempt to mitigate computation complexity by approximation algorithms. However, they need to retrain for each test sample, leading to intolerable costs. We propose Fast-DataShapley, a one-pass training method that leverages the weighted least squares characterization of the Shapley value to train a reusable explainer model with real-time reasoning speed. Given new test samples, no retraining is required to calculate the Shapley values of the training data. Additionally, we propose three methods with theoretical guarantees to reduce training overhead from two aspects: the approximate calculation of the utility function and the group calculation of the training data. We analyze time complexity to show the efficiency of our methods. The experimental evaluations on various image datasets demonstrate superior performance and efficiency compared to baselines. Specifically, the performance is improved to more than 2.5 times, and the explainer's training speed can be increased by two orders of magnitude. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:35:46 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Sun",
"Haifeng",
""
],
[
"Xiong",
"Yu",
""
],
[
"Wu",
"Runze",
""
],
[
"Cai",
"Xinyu",
""
],
[
"Fan",
"Changjie",
""
],
[
"Zhang",
"Lan",
""
],
[
"Li",
"Xiang-Yang",
""
]
] |
2506.05282 | Liyuan Zhu | Tao Sun and Liyuan Zhu and Shengyu Huang and Shuran Song and Iro Armeni | Rectified Point Flow: Generic Point Cloud Pose Estimation | Project page: https://rectified-pointflow.github.io/ | null | null | null | cs.CV cs.AI cs.RO | http://creativecommons.org/licenses/by/4.0/ | We introduce Rectified Point Flow, a unified parameterization that formulates pairwise point cloud registration and multi-part shape assembly as a single conditional generative problem. Given unposed point clouds, our method learns a continuous point-wise velocity field that transports noisy points toward their target positions, from which part poses are recovered. In contrast to prior work that regresses part-wise poses with ad-hoc symmetry handling, our method intrinsically learns assembly symmetries without symmetry labels. Together with a self-supervised encoder focused on overlapping points, our method achieves a new state-of-the-art performance on six benchmarks spanning pairwise registration and shape assembly. Notably, our unified formulation enables effective joint training on diverse datasets, facilitating the learning of shared geometric priors and consequently boosting accuracy. Project page: https://rectified-pointflow.github.io/. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:36:03 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Sun",
"Tao",
""
],
[
"Zhu",
"Liyuan",
""
],
[
"Huang",
"Shengyu",
""
],
[
"Song",
"Shuran",
""
],
[
"Armeni",
"Iro",
""
]
] |
2506.05283 | Michael Ruderman | Michael Ruderman, Denis Efimov | Regularization of non-overshooting quasi-continuous sliding mode control for chattering suppression at equilibrium | 6 pages, 2 figures | null | null | null | math.OC cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | Robust finite-time feedback controller introduced for the second-order systems in [1] can be seen as a non-overshooting quasi-continuous sliding mode control. The paper proposes a regularization scheme to suppress inherent chattering due to discontinuity of the control [1] in the origin, in favor of practical applications. A detailed analysis with ISS and iISS proofs are provided along with supporting numerical results. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:39:57 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ruderman",
"Michael",
""
],
[
"Efimov",
"Denis",
""
]
] |
2506.05284 | Tong Wu | Tong Wu, Shuai Yang, Ryan Po, Yinghao Xu, Ziwei Liu, Dahua Lin, Gordon Wetzstein | Video World Models with Long-term Spatial Memory | Project page: https://spmem.github.io/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Emerging world models autoregressively generate video frames in response to actions, such as camera movements and text prompts, among other control signals. Due to limited temporal context window sizes, these models often struggle to maintain scene consistency during revisits, leading to severe forgetting of previously generated environments. Inspired by the mechanisms of human memory, we introduce a novel framework to enhancing long-term consistency of video world models through a geometry-grounded long-term spatial memory. Our framework includes mechanisms to store and retrieve information from the long-term spatial memory and we curate custom datasets to train and evaluate world models with explicitly stored 3D memory mechanisms. Our evaluations show improved quality, consistency, and context length compared to relevant baselines, paving the way towards long-term consistent world generation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:42:34 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wu",
"Tong",
""
],
[
"Yang",
"Shuai",
""
],
[
"Po",
"Ryan",
""
],
[
"Xu",
"Yinghao",
""
],
[
"Liu",
"Ziwei",
""
],
[
"Lin",
"Dahua",
""
],
[
"Wetzstein",
"Gordon",
""
]
] |
2506.05285 | Bardienus Duisterhof | Bardienus P. Duisterhof, Jan Oberst, Bowen Wen, Stan Birchfield, Deva Ramanan, Jeffrey Ichnowski | RaySt3R: Predicting Novel Depth Maps for Zero-Shot Object Completion | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | 3D shape completion has broad applications in robotics, digital twin reconstruction, and extended reality (XR). Although recent advances in 3D object and scene completion have achieved impressive results, existing methods lack 3D consistency, are computationally expensive, and struggle to capture sharp object boundaries. Our work (RaySt3R) addresses these limitations by recasting 3D shape completion as a novel view synthesis problem. Specifically, given a single RGB-D image and a novel viewpoint (encoded as a collection of query rays), we train a feedforward transformer to predict depth maps, object masks, and per-pixel confidence scores for those query rays. RaySt3R fuses these predictions across multiple query views to reconstruct complete 3D shapes. We evaluate RaySt3R on synthetic and real-world datasets, and observe it achieves state-of-the-art performance, outperforming the baselines on all datasets by up to 44% in 3D chamfer distance. Project page: https://rayst3r.github.io | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:43:23 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Duisterhof",
"Bardienus P.",
""
],
[
"Oberst",
"Jan",
""
],
[
"Wen",
"Bowen",
""
],
[
"Birchfield",
"Stan",
""
],
[
"Ramanan",
"Deva",
""
],
[
"Ichnowski",
"Jeffrey",
""
]
] |
2506.05286 | Lijie Hu | Lijie Hu, Songning Lai, Yuan Hua, Shu Yang, Jingfeng Zhang, Di Wang | Stable Vision Concept Transformers for Medical Diagnosis | arXiv admin note: text overlap with arXiv:2304.06129 by other authors | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Transparency is a paramount concern in the medical field, prompting researchers to delve into the realm of explainable AI (XAI). Among these XAI methods, Concept Bottleneck Models (CBMs) aim to restrict the model's latent space to human-understandable high-level concepts by generating a conceptual layer for extracting conceptual features, which has drawn much attention recently. However, existing methods rely solely on concept features to determine the model's predictions, which overlook the intrinsic feature embeddings within medical images. To address this utility gap between the original models and concept-based models, we propose Vision Concept Transformer (VCT). Furthermore, despite their benefits, CBMs have been found to negatively impact model performance and fail to provide stable explanations when faced with input perturbations, which limits their application in the medical field. To address this faithfulness issue, this paper further proposes the Stable Vision Concept Transformer (SVCT) based on VCT, which leverages the vision transformer (ViT) as its backbone and incorporates a conceptual layer. SVCT employs conceptual features to enhance decision-making capabilities by fusing them with image features and ensures model faithfulness through the integration of Denoised Diffusion Smoothing. Comprehensive experiments on four medical datasets demonstrate that our VCT and SVCT maintain accuracy while remaining interpretable compared to baselines. Furthermore, even when subjected to perturbations, our SVCT model consistently provides faithful explanations, thus meeting the needs of the medical field. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:43:27 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Hu",
"Lijie",
""
],
[
"Lai",
"Songning",
""
],
[
"Hua",
"Yuan",
""
],
[
"Yang",
"Shu",
""
],
[
"Zhang",
"Jingfeng",
""
],
[
"Wang",
"Di",
""
]
] |
2506.05287 | Yuqian Yuan | Yuqian Yuan, Ronghao Dang, Long Li, Wentong Li, Dian Jiao, Xin Li, Deli Zhao, Fan Wang, Wenqiao Zhang, Jun Xiao, Yueting Zhuang | EOC-Bench: Can MLLMs Identify, Recall, and Forecast Objects in an Egocentric World? | 32pages | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The emergence of multimodal large language models (MLLMs) has driven breakthroughs in egocentric vision applications. These applications necessitate persistent, context-aware understanding of objects, as users interact with tools in dynamic and cluttered environments. However, existing embodied benchmarks primarily focus on static scene exploration, emphasizing object's appearance and spatial attributes while neglecting the assessment of dynamic changes arising from users' interactions. To address this gap, we introduce EOC-Bench, an innovative benchmark designed to systematically evaluate object-centric embodied cognition in dynamic egocentric scenarios. Specially, EOC-Bench features 3,277 meticulously annotated QA pairs categorized into three temporal categories: Past, Present, and Future, covering 11 fine-grained evaluation dimensions and 3 visual object referencing types. To ensure thorough assessment, we develop a mixed-format human-in-the-loop annotation framework with four types of questions and design a novel multi-scale temporal accuracy metric for open-ended temporal evaluation. Based on EOC-Bench, we conduct comprehensive evaluations of various proprietary, open-source, and object-level MLLMs. EOC-Bench serves as a crucial tool for advancing the embodied object cognitive capabilities of MLLMs, establishing a robust foundation for developing reliable core models for embodied systems. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:44:12 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Yuan",
"Yuqian",
""
],
[
"Dang",
"Ronghao",
""
],
[
"Li",
"Long",
""
],
[
"Li",
"Wentong",
""
],
[
"Jiao",
"Dian",
""
],
[
"Li",
"Xin",
""
],
[
"Zhao",
"Deli",
""
],
[
"Wang",
"Fan",
""
],
[
"Zhang",
"Wenqiao",
""
],
[
"Xiao",
"Jun",
""
],
[
"Zhuang",
"Yueting",
""
]
] |
2506.05288 | Philip Bittihn | Lukas Hupe, Joanna M. Materska, David Zwicker, Ramin Golestanian, Bartlomiej Waclaw, Philip Bittihn | Phase separation in a mixture of proliferating and motile active matter | 6 pages, 4 figures | null | null | null | cond-mat.soft cond-mat.stat-mech physics.bio-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Proliferation and motility are ubiquitous drivers of activity in biological systems. Here, we study a dense binary mixture of motile and proliferating particles with exclusively repulsive interactions, where homeostasis in the proliferating subpopulation is maintained by pressure-induced removal. Using computer simulations, we show that phase separation emerges naturally in this system at high density and weak enough self-propulsion. We show that condensation is caused by interactions between motile particles induced by the growing phase, and recapitulate this behavior in an effective model of only motile particles with attractive interactions. Our results establish a new type of phase transition and pave a way to reinterpret the physics of dense cellular populations, such as bacterial colonies or tumors, as systems of mixed active matter. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:44:24 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Hupe",
"Lukas",
""
],
[
"Materska",
"Joanna M.",
""
],
[
"Zwicker",
"David",
""
],
[
"Golestanian",
"Ramin",
""
],
[
"Waclaw",
"Bartlomiej",
""
],
[
"Bittihn",
"Philip",
""
]
] |
2506.05289 | Pingyu Wu | Pingyu Wu, Kai Zhu, Yu Liu, Longxiang Tang, Jian Yang, Yansong Peng, Wei Zhai, Yang Cao, Zheng-Jun Zha | AliTok: Towards Sequence Modeling Alignment between Tokenizer and Autoregressive Model | Code: https://github.com/ali-vilab/alitok | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autoregressive image generation aims to predict the next token based on previous ones. However, existing image tokenizers encode tokens with bidirectional dependencies during the compression process, which hinders the effective modeling by autoregressive models. In this paper, we propose a novel Aligned Tokenizer (AliTok), which utilizes a causal decoder to establish unidirectional dependencies among encoded tokens, thereby aligning the token modeling approach between the tokenizer and autoregressive model. Furthermore, by incorporating prefix tokens and employing two-stage tokenizer training to enhance reconstruction consistency, AliTok achieves great reconstruction performance while being generation-friendly. On ImageNet-256 benchmark, using a standard decoder-only autoregressive model as the generator with only 177M parameters, AliTok achieves a gFID score of 1.50 and an IS of 305.9. When the parameter count is increased to 662M, AliTok achieves a gFID score of 1.35, surpassing the state-of-the-art diffusion method with 10x faster sampling speed. The code and weights are available at https://github.com/ali-vilab/alitok. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:45:10 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wu",
"Pingyu",
""
],
[
"Zhu",
"Kai",
""
],
[
"Liu",
"Yu",
""
],
[
"Tang",
"Longxiang",
""
],
[
"Yang",
"Jian",
""
],
[
"Peng",
"Yansong",
""
],
[
"Zhai",
"Wei",
""
],
[
"Cao",
"Yang",
""
],
[
"Zha",
"Zheng-Jun",
""
]
] |
2506.05290 | Pierre Tholoniat | Pierre Tholoniat, Alison Caulfield, Giorgio Cavicchioli, Mark Chen, Nikos Goutzoulias, Benjamin Case, Asaf Cidon, Roxana Geambasu, Mathias L\'ecuyer, Martin Thomson | Big Bird: Privacy Budget Management for W3C's Privacy-Preserving Attribution API | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by-sa/4.0/ | Privacy-preserving advertising APIs like Privacy-Preserving Attribution (PPA) are designed to enhance web privacy while enabling effective ad measurement. PPA offers an alternative to cross-site tracking with encrypted reports governed by differential privacy (DP), but current designs lack a principled approach to privacy budget management, creating uncertainty around critical design decisions. We present Big Bird, a privacy budget manager for PPA that clarifies per-site budget semantics and introduces a global budgeting system grounded in resource isolation principles. Big Bird enforces utility-preserving limits via quota budgets and improves global budget utilization through a novel batched scheduling algorithm. Together, these mechanisms establish a robust foundation for enforcing privacy protections in adversarial environments. We implement Big Bird in Firefox and evaluate it on real-world ad data, demonstrating its resilience and effectiveness. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:45:13 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Tholoniat",
"Pierre",
""
],
[
"Caulfield",
"Alison",
""
],
[
"Cavicchioli",
"Giorgio",
""
],
[
"Chen",
"Mark",
""
],
[
"Goutzoulias",
"Nikos",
""
],
[
"Case",
"Benjamin",
""
],
[
"Cidon",
"Asaf",
""
],
[
"Geambasu",
"Roxana",
""
],
[
"Lécuyer",
"Mathias",
""
],
[
"Thomson",
"Martin",
""
]
] |
2506.05291 | Yu Jiang | Yu Jiang | On elementary abelian 2-hypergroups | 18 pages | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A hypergroup is called an elementary abelian 2-hypergroup if it is a constrained direct product of the closed subsets of two elements. In this paper, the elementary abelian 2-hypergroups are studied. All closed subsets and all strongly normal closed subsets of the elementary abelian 2-hypergroups are determined. The numbers of all closed subsets and all strongly normal closed subsets of the elementary abelian 2-hypergroups are given. A criterion for the isomorphic closed subsets of the elementary abelian 2-hypergroups is displayed. The automorphism groups of all closed subsets of the elementary abelian 2-hypergroups are presented. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:45:16 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Jiang",
"Yu",
""
]
] |
2506.05292 | Declan Alex Norton | Declan A. Norton, Yuanzhao Zhang, Michelle Girvan | Learning Beyond Experience: Generalizing to Unseen State Space with Reservoir Computing | 15 pages, 9 figures | null | null | null | cs.LG math.DS nlin.CD physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | Machine learning techniques offer an effective approach to modeling dynamical systems solely from observed data. However, without explicit structural priors -- built-in assumptions about the underlying dynamics -- these techniques typically struggle to generalize to aspects of the dynamics that are poorly represented in the training data. Here, we demonstrate that reservoir computing -- a simple, efficient, and versatile machine learning framework often used for data-driven modeling of dynamical systems -- can generalize to unexplored regions of state space without explicit structural priors. First, we describe a multiple-trajectory training scheme for reservoir computers that supports training across a collection of disjoint time series, enabling effective use of available training data. Then, applying this training scheme to multistable dynamical systems, we show that RCs trained on trajectories from a single basin of attraction can achieve out-of-domain generalization by capturing system behavior in entirely unobserved basins. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:46:07 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Norton",
"Declan A.",
""
],
[
"Zhang",
"Yuanzhao",
""
],
[
"Girvan",
"Michelle",
""
]
] |
2506.05293 | Mario Christopher Bedrunka | Mario Christopher Bedrunka, Dirk Reith, Holger Foysi, {\L}ukasz {\L}aniewski-Wo{\l}{\l}k and Travis Mitchell | Reduction of Outflow Boundary Influence on Aerodynamic Performance using Neural Networks | null | null | null | null | physics.comp-ph physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | The accurate treatment of outflow boundary conditions remains a critical challenge in computational fluid dynamics when predicting aerodynamic forces and/or acoustic emissions. This is particularly evident when employing the lattice Boltzmann method (LBM) as the numerical solution technique, which often suffers from inaccuracies induced by artificial reflections from outflow boundaries. This paper investigates the use of neural networks (NN) to mitigate these adverse boundary effects and enable truncated domain requirements. Two distinct NN-based approaches are proposed: (1) direct reconstruction of unknown particle distribution functions at the outflow boundary; and (2) enhancement of established characteristic boundary conditions (CBC) by dynamically tuning their parameters. The direct reconstruction model was trained on data generated from a 2D flow over a cylindrical obstruction. The drag, lift, and Strouhal number were used to test the new boundary condition. We analyzed results for various Reynolds numbers and restricted domain sizes where it demonstrated significantly improved predictions when compared with the traditional Zou & He boundary condition. To examine the robustness of the NN-based reconstruction, the same condition was applied to the simulation of a NACA0012 airfoil, again providing accurate aerodynamic performance predictions. The neural-enhanced CBC were evaluated on a 2D convected vortex benchmark and showed superior performance in minimizing density errors compared to CBCs with fixed parameters. These findings highlight the potential of NN-integrated boundary conditions to improve accuracy and reduce computational expense of aerodynamic and acoustic emissions simulations with the LBM. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:47:03 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Bedrunka",
"Mario Christopher",
""
],
[
"Reith",
"Dirk",
""
],
[
"Foysi",
"Holger",
""
],
[
"Łaniewski-Wołłk",
"Łukasz",
""
],
[
"Mitchell",
"Travis",
""
]
] |
2506.05294 | Arnav Kumar Jain | Arnav Kumar Jain, Vibhakar Mohta, Subin Kim, Atiksh Bhardwaj, Juntao Ren, Yunhai Feng, Sanjiban Choudhury, Gokul Swamy | A Smooth Sea Never Made a Skilled $\texttt{SAILOR}$: Robust Imitation via Learning to Search | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | The fundamental limitation of the behavioral cloning (BC) approach to imitation learning is that it only teaches an agent what the expert did at states the expert visited. This means that when a BC agent makes a mistake which takes them out of the support of the demonstrations, they often don't know how to recover from it. In this sense, BC is akin to giving the agent the fish -- giving them dense supervision across a narrow set of states -- rather than teaching them to fish: to be able to reason independently about achieving the expert's outcome even when faced with unseen situations at test-time. In response, we explore learning to search (L2S) from expert demonstrations, i.e. learning the components required to, at test time, plan to match expert outcomes, even after making a mistake. These include (1) a world model and (2) a reward model. We carefully ablate the set of algorithmic and design decisions required to combine these and other components for stable and sample/interaction-efficient learning of recovery behavior without additional human corrections. Across a dozen visual manipulation tasks from three benchmarks, our approach $\texttt{SAILOR}$ consistently out-performs state-of-the-art Diffusion Policies trained via BC on the same data. Furthermore, scaling up the amount of demonstrations used for BC by 5-10$\times$ still leaves a performance gap. We find that $\texttt{SAILOR}$ can identify nuanced failures and is robust to reward hacking. Our code is available at https://github.com/arnavkj1995/SAILOR . | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:47:40 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Jain",
"Arnav Kumar",
""
],
[
"Mohta",
"Vibhakar",
""
],
[
"Kim",
"Subin",
""
],
[
"Bhardwaj",
"Atiksh",
""
],
[
"Ren",
"Juntao",
""
],
[
"Feng",
"Yunhai",
""
],
[
"Choudhury",
"Sanjiban",
""
],
[
"Swamy",
"Gokul",
""
]
] |
2506.05295 | Baihe Huang | Baihe Huang, Shanda Li, Tianhao Wu, Yiming Yang, Ameet Talwalkar, Kannan Ramchandran, Michael I. Jordan, Jiantao Jiao | Sample Complexity and Representation Ability of Test-time Scaling Paradigms | null | null | null | null | cs.LG cs.AI stat.ML | http://creativecommons.org/licenses/by/4.0/ | Test-time scaling paradigms have significantly advanced the capabilities of large language models (LLMs) on complex tasks. Despite their empirical success, theoretical understanding of the sample efficiency of various test-time strategies -- such as self-consistency, best-of-$n$, and self-correction -- remains limited. In this work, we first establish a separation result between two repeated sampling strategies: self-consistency requires $\Theta(1/\Delta^2)$ samples to produce the correct answer, while best-of-$n$ only needs $\Theta(1/\Delta)$, where $\Delta < 1$ denotes the probability gap between the correct and second most likely answers. Next, we present an expressiveness result for the self-correction approach with verifier feedback: it enables Transformers to simulate online learning over a pool of experts at test time. Therefore, a single Transformer architecture can provably solve multiple tasks without prior knowledge of the specific task associated with a user query, extending the representation theory of Transformers from single-task to multi-task settings. Finally, we empirically validate our theoretical results, demonstrating the practical effectiveness of self-correction methods. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:48:19 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Huang",
"Baihe",
""
],
[
"Li",
"Shanda",
""
],
[
"Wu",
"Tianhao",
""
],
[
"Yang",
"Yiming",
""
],
[
"Talwalkar",
"Ameet",
""
],
[
"Ramchandran",
"Kannan",
""
],
[
"Jordan",
"Michael I.",
""
],
[
"Jiao",
"Jiantao",
""
]
] |
2506.05296 | Mikhail Terekhov | Mikhail Terekhov, Zhen Ning David Liu, Caglar Gulcehre, Samuel Albanie | Control Tax: The Price of Keeping AI in Check | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | The rapid integration of agentic AI into high-stakes real-world applications requires robust oversight mechanisms. The emerging field of AI Control (AIC) aims to provide such an oversight mechanism, but practical adoption depends heavily on implementation overhead. To study this problem better, we introduce the notion of Control tax -- the operational and financial cost of integrating control measures into AI pipelines. Our work makes three key contributions to the field of AIC: (1) we introduce a theoretical framework that quantifies the Control Tax and maps classifier performance to safety assurances; (2) we conduct comprehensive evaluations of state-of-the-art language models in adversarial settings, where attacker models insert subtle backdoors into code while monitoring models attempt to detect these vulnerabilities; and (3) we provide empirical financial cost estimates for control protocols and develop optimized monitoring strategies that balance safety and cost-effectiveness while accounting for practical constraints like auditing budgets. Our framework enables practitioners to make informed decisions by systematically connecting safety guarantees with their costs, advancing AIC through principled economic feasibility assessment across different deployment contexts. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:48:39 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Terekhov",
"Mikhail",
""
],
[
"Liu",
"Zhen Ning David",
""
],
[
"Gulcehre",
"Caglar",
""
],
[
"Albanie",
"Samuel",
""
]
] |
2506.05297 | Hangyu Ji | Hangyu Ji | DM-SegNet: Dual-Mamba Architecture for 3D Medical Image Segmentation with Global Context Modeling | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate 3D medical image segmentation demands architectures capable of reconciling global context modeling with spatial topology preservation. While State Space Models (SSMs) like Mamba show potential for sequence modeling, existing medical SSMs suffer from encoder-decoder incompatibility: the encoder's 1D sequence flattening compromises spatial structures, while conventional decoders fail to leverage Mamba's state propagation. We present DM-SegNet, a Dual-Mamba architecture integrating directional state transitions with anatomy-aware hierarchical decoding. The core innovations include a quadri-directional spatial Mamba module employing four-directional 3D scanning to maintain anatomical spatial coherence, a gated spatial convolution layer that enhances spatially sensitive feature representation prior to state modeling, and a Mamba-driven decoding framework enabling bidirectional state synchronization across scales. Extensive evaluation on two clinically significant benchmarks demonstrates the efficacy of DM-SegNet: achieving state-of-the-art Dice Similarity Coefficient (DSC) of 85.44% on the Synapse dataset for abdominal organ segmentation and 90.22% on the BraTS2023 dataset for brain tumor segmentation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:49:46 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ji",
"Hangyu",
""
]
] |
2506.05298 | Alexandru Lupsasca | Alexandru Lupsasca | Why there is no Love in black holes | 13 pages, no figures | null | null | null | gr-qc hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a new conformal symmetry of stationary, axisymmetric Kerr perturbations. This symmetry is exact but non-geometric (or "hidden"), and each of its generators has an associated infinite family of eigenstate solutions. Tidal perturbations of a black hole form an irreducible highest-weight representation of this conformal group, while the tidal response fields live in a different such representation. This implies that black holes have no tidal deformability, or vanishing Love numbers. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:50:18 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Lupsasca",
"Alexandru",
""
]
] |
2506.05299 | Christos Sourdis | Christos Sourdis | On a linear equation arising in the study of phase separation of BEC's | null | null | null | null | math.AP | http://creativecommons.org/publicdomain/zero/1.0/ | We consider the inner limit system describing the phase separation in two-component Bose-Einstein condensates linearized around the one-dimensional solution in an infinite strip with zero and periodic boundary conditions, and obtain optimal invertibility estimates for the Fourier modes without necessarily assuming orthogonality conditions. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:50:28 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Sourdis",
"Christos",
""
]
] |
2506.05300 | Prajwal Singhania | Nirav Koley, Prajwal Singhania, Abhinav Bhatele | Power Law Guided Dynamic Sifting for Efficient Attention | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Efficient inference on GPUs using large language models remains challenging due to memory bandwidth limitations, particularly during data transfers between High Bandwidth Memory (HBM) and SRAM in attention computations. Approximate attention methods address this issue by reducing computational and memory overhead but often rely on expensive top-$k$ operations, which perform poorly on GPUs. We propose SiftAttention, a novel approximate attention method that replaces the top-$k$ step with a computationally efficient element-wise filtering operation based on a threshold value. Our intuition for doing this is based on our empirical observation that the $\tau$-th quantile of attention scores follows a predictable power-law over sequential generation steps. Exploiting this insight, our approach dynamically estimates a threshold value per prompt at each generation step. Only attention scores above this threshold and their corresponding value vectors are loaded/used to compute the attention output, reducing data movement between HBM and SRAM. Our evaluation demonstrates that SiftAttention preserves model quality better than existing approximate attention methods while reducing memory bandwidth usage when loading value vectors. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:50:32 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Koley",
"Nirav",
""
],
[
"Singhania",
"Prajwal",
""
],
[
"Bhatele",
"Abhinav",
""
]
] |
2506.05301 | Jianyi Wang | Jianyi Wang, Shanchuan Lin, Zhijie Lin, Yuxi Ren, Meng Wei, Zongsheng Yue, Shangchen Zhou, Hao Chen, Yang Zhao, Ceyuan Yang, Xuefeng Xiao, Chen Change Loy, Lu Jiang | SeedVR2: One-Step Video Restoration via Diffusion Adversarial Post-Training | Draft Ver. Project page: https://iceclear.github.io/projects/seedvr2/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in diffusion-based video restoration (VR) demonstrate significant improvement in visual quality, yet yield a prohibitive computational cost during inference. While several distillation-based approaches have exhibited the potential of one-step image restoration, extending existing approaches to VR remains challenging and underexplored, particularly when dealing with high-resolution video in real-world settings. In this work, we propose a one-step diffusion-based VR model, termed as SeedVR2, which performs adversarial VR training against real data. To handle the challenging high-resolution VR within a single step, we introduce several enhancements to both model architecture and training procedures. Specifically, an adaptive window attention mechanism is proposed, where the window size is dynamically adjusted to fit the output resolutions, avoiding window inconsistency observed under high-resolution VR using window attention with a predefined window size. To stabilize and improve the adversarial post-training towards VR, we further verify the effectiveness of a series of losses, including a proposed feature matching loss without significantly sacrificing training efficiency. Extensive experiments show that SeedVR2 can achieve comparable or even better performance compared with existing VR approaches in a single step. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:51:05 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wang",
"Jianyi",
""
],
[
"Lin",
"Shanchuan",
""
],
[
"Lin",
"Zhijie",
""
],
[
"Ren",
"Yuxi",
""
],
[
"Wei",
"Meng",
""
],
[
"Yue",
"Zongsheng",
""
],
[
"Zhou",
"Shangchen",
""
],
[
"Chen",
"Hao",
""
],
[
"Zhao",
"Yang",
""
],
[
"Yang",
"Ceyuan",
""
],
[
"Xiao",
"Xuefeng",
""
],
[
"Loy",
"Chen Change",
""
],
[
"Jiang",
"Lu",
""
]
] |
2506.05302 | Weifeng Lin | Weifeng Lin, Xinyu Wei, Ruichuan An, Tianhe Ren, Tingwei Chen, Renrui Zhang, Ziyu Guo, Wentao Zhang, Lei Zhang and Hongsheng Li | Perceive Anything: Recognize, Explain, Caption, and Segment Anything in Images and Videos | 19 pages, 13 figures, Website: https://Perceive-Anything.github.io | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We present Perceive Anything Model (PAM), a conceptually straightforward and efficient framework for comprehensive region-level visual understanding in images and videos. Our approach extends the powerful segmentation model SAM 2 by integrating Large Language Models (LLMs), enabling simultaneous object segmentation with the generation of diverse, region-specific semantic outputs, including categories, label definition, functional explanations, and detailed captions. A key component, Semantic Perceiver, is introduced to efficiently transform SAM 2's rich visual features, which inherently carry general vision, localization, and semantic priors into multi-modal tokens for LLM comprehension. To support robust multi-granularity understanding, we also develop a dedicated data refinement and augmentation pipeline, yielding a high-quality dataset of 1.5M image and 0.6M video region-semantic annotations, including novel region-level streaming video caption data. PAM is designed for lightweightness and efficiency, while also demonstrates strong performance across a diverse range of region understanding tasks. It runs 1.2-2.4x faster and consumes less GPU memory than prior approaches, offering a practical solution for real-world applications. We believe that our effective approach will serve as a strong baseline for future research in region-level visual understanding. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:51:39 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Lin",
"Weifeng",
""
],
[
"Wei",
"Xinyu",
""
],
[
"An",
"Ruichuan",
""
],
[
"Ren",
"Tianhe",
""
],
[
"Chen",
"Tingwei",
""
],
[
"Zhang",
"Renrui",
""
],
[
"Guo",
"Ziyu",
""
],
[
"Zhang",
"Wentao",
""
],
[
"Zhang",
"Lei",
""
],
[
"Li",
"Hongsheng",
""
]
] |
2506.05303 | David Clark | David G. Clark | Transient dynamics of associative memory models | 18 pages, 7 figures | null | null | null | cond-mat.dis-nn q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Associative memory models such as the Hopfield network and its dense generalizations with higher-order interactions exhibit a "blackout catastrophe"--a discontinuous transition where stable memory states abruptly vanish when the number of stored patterns exceeds a critical capacity. This transition is often interpreted as rendering networks unusable beyond capacity limits. We argue that this interpretation is largely an artifact of the equilibrium perspective. We derive dynamical mean-field equations using a bipartite cavity approach for graded-activity dense associative memory models, with the Hopfield model as a special case, and solve them using a numerical scheme. We show that patterns can be transiently retrieved with high accuracy above capacity despite the absence of stable attractors. This occurs because slow regions persist in the above-capacity energy landscape as shallow, unstable remnants of below-capacity stable basins. The same transient-retrieval effect occurs in below-capacity networks initialized outside basins of attraction. "Transient-recovery curves" provide a concise visual summary of these effects, revealing graceful, non-catastrophic changes in retrieval behavior above capacity and allowing us to compare the behavior across interaction orders. This dynamical perspective reveals rich energy landscape structure obscured by equilibrium analysis and suggests biological neural circuits may exploit transient dynamics for memory retrieval. Furthermore, our approach suggests ways of understanding computational properties of neural circuits without reference to fixed points, advances the technical repertoire of numerical mean-field solution methods for recurrent neural networks, and yields new theoretical results on generalizations of the Hopfield model. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:51:39 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Clark",
"David G.",
""
]
] |
2506.05304 | Youssef Hassan | Youssef S. Hassan, Kyle Beloy, Jacob L. Siegel, Takumi Kobayashi, Eric Swiler, Tanner Grogan, Roger C. Brown, Tristan Rojo, Tobias Bothwell, Benjamin D. Hunt, Adam Halaoui and Andrew D. Ludlow | Cryogenic Optical Lattice Clock with $1.7\times 10^{-20}$ Blackbody Radiation Stark Uncertainty | 21 pages (7 main + 14 SM), 6 figures (3 main + 3 SM), 1 table, to be published in Physical Review Letters | null | null | null | physics.atom-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Controlling the Stark perturbation from ambient thermal radiation is key to advancing the performance of many atomic frequency standards, including state-of-the-art optical lattice clocks (OLCs). We demonstrate a cryogenic OLC that utilizes a dynamically actuated radiation shield to control the perturbation at $1.7\times10^{-20}$ fractional frequency, a factor of $\sim$40 beyond the best OLC to date. Our shield furnishes the atoms with a near-ideal cryogenic blackbody radiation (BBR) environment by rejecting external thermal radiation at the part-per-million level during clock spectroscopy, overcoming a key limitation with previous cryogenic BBR control solutions in OLCs. While the lowest BBR shift uncertainty is realized with cryogenic operation, we further exploit the radiation control that the shield offers over a wide range of temperatures to directly measure and verify the leading BBR Stark dynamic correction coefficient for ytterbium. This independent measurement reduces the literature-combined uncertainty of this coefficient by 30%, thus benefiting state-of-the-art Yb OLCs operated at room temperature. We verify the static BBR coefficient for Yb at the low $10^{-18}$ level. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:51:44 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Hassan",
"Youssef S.",
""
],
[
"Beloy",
"Kyle",
""
],
[
"Siegel",
"Jacob L.",
""
],
[
"Kobayashi",
"Takumi",
""
],
[
"Swiler",
"Eric",
""
],
[
"Grogan",
"Tanner",
""
],
[
"Brown",
"Roger C.",
""
],
[
"Rojo",
"Tristan",
""
],
[
"Bothwell",
"Tobias",
""
],
[
"Hunt",
"Benjamin D.",
""
],
[
"Halaoui",
"Adam",
""
],
[
"Ludlow",
"Andrew D.",
""
]
] |
2506.05305 | Deepak Pandita | Deepak Pandita, Tharindu Cyril Weerasooriya, Ankit Parag Shah, Christopher M. Homan, Wei Wei | ProRefine: Inference-time Prompt Refinement with Textual Feedback | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Agentic workflows, where multiple AI agents collaborate to accomplish complex tasks like reasoning or planning, are becoming increasingly prevalent. However, these workflows often suffer from error propagation and sub-optimal performance, largely due to poorly designed prompts that fail to effectively guide individual agents. This is a critical problem because it limits the reliability and scalability of these powerful systems. We introduce ProRefine, an innovative inference-time prompt optimization method that leverages textual feedback from large language models (LLMs) to address this challenge. ProRefine dynamically refines prompts for multi-step reasoning tasks without additional training or ground truth labels. Evaluated on five benchmark mathematical reasoning datasets, ProRefine significantly surpasses zero-shot Chain-of-Thought baselines by 3 to 37 percentage points. This approach not only boosts accuracy but also allows smaller models to match the performance of larger ones, highlighting its potential for efficient and scalable AI deployment, and democratizing access to high-performing AI. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:52:30 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Pandita",
"Deepak",
""
],
[
"Weerasooriya",
"Tharindu Cyril",
""
],
[
"Shah",
"Ankit Parag",
""
],
[
"Homan",
"Christopher M.",
""
],
[
"Wei",
"Wei",
""
]
] |
2506.05306 | Pavel Kurilovich | Thomas Connolly, Pavel D. Kurilovich, Vladislav D. Kurilovich, Charlotte G. L. B{\o}ttcher, Sumeru Hazra, Wei Dai, Andy Z. Ding, Vidul R. Joshi, Heekun Nho, Spencer Diamond, Daniel K. Weiss, Valla Fatemi, Luigi Frunzio, Leonid I. Glazman, and Michel H. Devoret | Full characterization of measurement-induced transitions of a superconducting qubit | 30 pages, 16 figures | null | null | null | quant-ph cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | Repeated quantum non-demolition measurement is a cornerstone of quantum error correction protocols. In superconducting qubits, the speed of dispersive state readout can be enhanced by increasing the power of the readout tone. However, such an increase has been found to result in additional qubit state transitions that violate the desired quantum non-demolition character of the measurement. Recently, the readout of a transmon superconducting qubit was improved by using a tone with frequency much larger than the qubit frequency. Here, we experimentally identify the mechanisms of readout-induced transitions in this regime. In the dominant mechanism, the energy of an incoming readout photon is partially absorbed by the transmon and partially returned to the transmission line as a photon with lower frequency. Other mechanisms involve the excitation of unwanted package modes, decay via material defects, and, at higher qubit frequencies, the activation of undesired resonances in the transmon spectrum. Our work provides a comprehensive characterization of superconducting qubit state transitions caused by a strong drive. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:53:05 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Connolly",
"Thomas",
""
],
[
"Kurilovich",
"Pavel D.",
""
],
[
"Kurilovich",
"Vladislav D.",
""
],
[
"Bøttcher",
"Charlotte G. L.",
""
],
[
"Hazra",
"Sumeru",
""
],
[
"Dai",
"Wei",
""
],
[
"Ding",
"Andy Z.",
""
],
[
"Joshi",
"Vidul R.",
""
],
[
"Nho",
"Heekun",
""
],
[
"Diamond",
"Spencer",
""
],
[
"Weiss",
"Daniel K.",
""
],
[
"Fatemi",
"Valla",
""
],
[
"Frunzio",
"Luigi",
""
],
[
"Glazman",
"Leonid I.",
""
],
[
"Devoret",
"Michel H.",
""
]
] |
2506.05307 | Swati Choudhary | Himanshu Badhani, Dhanuja G S, Swati Choudhary, Vishal Anand, Siddhartha Das | Erasure cost of a quantum process: A thermodynamic meaning of the dynamical min-entropy | 8+9 pages, 3 figures | null | null | null | quant-ph math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The erasure of information is fundamentally an irreversible logical operation, carrying profound consequences for the energetics of computation and information processing. In this work, we investigate the thermodynamic costs associated with erasing (and preparing) quantum processes. Specifically, we analyze an arbitrary bipartite unitary gate acting on logical and ancillary input-output systems, where the ancillary input is always initialized in the ground state. We focus on the adversarial erasure cost of the reduced dynamics~\textemdash~that is, the minimal thermodynamic work required to erase the logical output of the gate for any logical input, assuming full access to the ancilla but no access to any purifying reference of the logical input state. We determine that this adversarial erasure cost is directly proportional to the negative min-entropy of the reduced dynamics, thereby giving the dynamical min-entropy a clear operational meaning. A key foundation of this result is the quantum process decoupling theorem, which quantitatively relates the decoupling ability of a process with its min-entropy. This insight bridges thermodynamics, information theory, and the fundamental limits of quantum computation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:53:17 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Badhani",
"Himanshu",
""
],
[
"S",
"Dhanuja G",
""
],
[
"Choudhary",
"Swati",
""
],
[
"Anand",
"Vishal",
""
],
[
"Das",
"Siddhartha",
""
]
] |
2506.05308 | Wenqian Xing | Ramesh Johari, Tianyi Peng, Wenqian Xing | Estimation of Treatment Effects Under Nonstationarity via Truncated Difference-in-Q's | 35 pages, 6 figures | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Randomized controlled experiments (''A/B testing'') are fundamental for assessing interventions in dynamic technology-driven environments, such as recommendation systems, online marketplaces, and digital health interventions. In these systems, interventions typically impact not only the current state of the system, but also future states; therefore, accurate estimation of the global average treatment effect (or GATE) from experiments requires accounting for the dynamic temporal behavior of the system. To address this, recent literature has analyzed a range of estimators applied to Bernoulli randomized experiments in stationary environments, ranging from the standard difference-in-means (DM) estimator to methods building on reinforcement learning techniques, such as off-policy evaluation and the recently proposed difference-in-Q's (DQ) estimator. However, all these estimators exhibit high bias and variance when the environment is nonstationary. This paper addresses the challenge of estimation under nonstationarity. We show that a simple extension of the DM estimator using differences in truncated outcome trajectories yields favorable bias and variance in nonstationary Markovian settings. Our theoretical analysis establishes this result by first showing that the truncated estimator is in fact estimating an appropriate policy gradient that can be expressed as a difference in Q-values; thus we refer to our estimator as the truncated DQ estimator (by analogy to the DQ estimator). We then show that the corresponding policy gradient is a first-order approximation to the GATE. Combining these insights yields our bias and variance bounds. We validate our results through synthetic and realistic simulations-including hospital and ride-sharing settings-and show that a well-calibrated truncated DQ estimator achieves low bias and variance even in nonstationary environments. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:53:26 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Johari",
"Ramesh",
""
],
[
"Peng",
"Tianyi",
""
],
[
"Xing",
"Wenqian",
""
]
] |
2506.05309 | Niv Eckhaus | Niv Eckhaus, Uri Berger and Gabriel Stanovsky | Time to Talk: LLM Agents for Asynchronous Group Communication in Mafia Games | null | null | null | null | cs.MA cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | LLMs are used predominantly in synchronous communication, where a human user and a model communicate in alternating turns. In contrast, many real-world settings are inherently asynchronous. For example, in group chats, online team meetings, or social games, there is no inherent notion of turns; therefore, the decision of when to speak forms a crucial part of the participant's decision making. In this work, we develop an adaptive asynchronous LLM-agent which, in addition to determining what to say, also decides when to say it. To evaluate our agent, we collect a unique dataset of online Mafia games, including both human participants, as well as our asynchronous agent. Overall, our agent performs on par with human players, both in game performance, as well as in its ability to blend in with the other human players. Our analysis shows that the agent's behavior in deciding when to speak closely mirrors human patterns, although differences emerge in message content. We release all our data and code to support and encourage further research for more realistic asynchronous communication between LLM agents. This work paves the way for integration of LLMs into realistic human group settings, from assistance in team discussions to educational and professional environments where complex social dynamics must be navigated. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:53:44 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Eckhaus",
"Niv",
""
],
[
"Berger",
"Uri",
""
],
[
"Stanovsky",
"Gabriel",
""
]
] |
2506.05310 | Florentin Guth | Florentin Guth, Zahra Kadkhodaie, Eero P Simoncelli | Learning normalized image densities via dual score matching | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Learning probability models from data is at the heart of many machine learning endeavors, but is notoriously difficult due to the curse of dimensionality. We introduce a new framework for learning \emph{normalized} energy (log probability) models that is inspired from diffusion generative models, which rely on networks optimized to estimate the score. We modify a score network architecture to compute an energy while preserving its inductive biases. The gradient of this energy network with respect to its input image is the score of the learned density, which can be optimized using a denoising objective. Importantly, the gradient with respect to the noise level provides an additional score that can be optimized with a novel secondary objective, ensuring consistent and normalized energies across noise levels. We train an energy network with this \emph{dual} score matching objective on the ImageNet64 dataset, and obtain a cross-entropy (negative log likelihood) value comparable to the state of the art. We further validate our approach by showing that our energy model \emph{strongly generalizes}: estimated log probabilities are nearly independent of the specific images in the training set. Finally, we demonstrate that both image probability and dimensionality of local neighborhoods vary significantly with image content, in contrast with traditional assumptions such as concentration of measure or support on a low-dimensional manifold. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:53:57 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Guth",
"Florentin",
""
],
[
"Kadkhodaie",
"Zahra",
""
],
[
"Simoncelli",
"Eero P",
""
]
] |
2506.05311 | Leandro Rizzi G. | T. N. Azevedo and L. G. Rizzi | Heterogeneous response and non-Markovianity in the microrheology of semisolid viscoelastic materials | 9 pages, 4 figures | null | null | null | cond-mat.soft cond-mat.mtrl-sci cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent works indicate that heterogeneous response and non-Markovianity may yield recognizable hallmarks in the microrheology of semisolid viscoelastic materials. Here we perform numerical simulations using a non-Markovian overdamped Langevin approach to explore how the microrheology experienced by probe particles immersed in an effective semisolid material can be influenced by its micro-heterogeneities. Our results show that, besides affecting the mean squared displacement, the time-dependent diffusion coefficient, and the shear moduli, the micro-heterogeneities lead to displacement distributions that deviate from the usual Gaussian behavior. In addition, our study provides an analytical way to characterize the micro-heterogeneities of semisolid viscoelastic materials through their microrheology. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:54:31 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Azevedo",
"T. N.",
""
],
[
"Rizzi",
"L. G.",
""
]
] |
2506.05312 | Olaf D\"unkel | Olaf D\"unkel, Thomas Wimmer, Christian Theobalt, Christian Rupprecht, Adam Kortylewski | Do It Yourself: Learning Semantic Correspondence from Pseudo-Labels | Project page: https://genintel.github.io/DIY-SC | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Finding correspondences between semantically similar points across images and object instances is one of the everlasting challenges in computer vision. While large pre-trained vision models have recently been demonstrated as effective priors for semantic matching, they still suffer from ambiguities for symmetric objects or repeated object parts. We propose to improve semantic correspondence estimation via 3D-aware pseudo-labeling. Specifically, we train an adapter to refine off-the-shelf features using pseudo-labels obtained via 3D-aware chaining, filtering wrong labels through relaxed cyclic consistency, and 3D spherical prototype mapping constraints. While reducing the need for dataset specific annotations compared to prior work, we set a new state-of-the-art on SPair-71k by over 4% absolute gain and by over 7% against methods with similar supervision requirements. The generality of our proposed approach simplifies extension of training to other data sources, which we demonstrate in our experiments. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:54:33 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Dünkel",
"Olaf",
""
],
[
"Wimmer",
"Thomas",
""
],
[
"Theobalt",
"Christian",
""
],
[
"Rupprecht",
"Christian",
""
],
[
"Kortylewski",
"Adam",
""
]
] |
2506.05313 | Ta-Ying Cheng | Ta-Ying Cheng, Prafull Sharma, Mark Boss, Varun Jampani | MARBLE: Material Recomposition and Blending in CLIP-Space | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Editing materials of objects in images based on exemplar images is an active area of research in computer vision and graphics. We propose MARBLE, a method for performing material blending and recomposing fine-grained material properties by finding material embeddings in CLIP-space and using that to control pre-trained text-to-image models. We improve exemplar-based material editing by finding a block in the denoising UNet responsible for material attribution. Given two material exemplar-images, we find directions in the CLIP-space for blending the materials. Further, we can achieve parametric control over fine-grained material attributes such as roughness, metallic, transparency, and glow using a shallow network to predict the direction for the desired material attribute change. We perform qualitative and quantitative analysis to demonstrate the efficacy of our proposed method. We also present the ability of our method to perform multiple edits in a single forward pass and applicability to painting.
Project Page: https://marblecontrol.github.io/ | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:55:16 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Cheng",
"Ta-Ying",
""
],
[
"Sharma",
"Prafull",
""
],
[
"Boss",
"Mark",
""
],
[
"Jampani",
"Varun",
""
]
] |
2506.05314 | Taha Entesari | Taha Entesari, Arman Hatami, Rinat Khaziev, Anil Ramakrishna, Mahyar Fazlyab | Constrained Entropic Unlearning: A Primal-Dual Framework for Large Language Models | null | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) deployed in real-world settings increasingly face the need to unlearn sensitive, outdated, or proprietary information. Existing unlearning methods typically formulate forgetting and retention as a regularized trade-off, combining both objectives into a single scalarized loss. This often leads to unstable optimization and degraded performance on retained data, especially under aggressive forgetting. We propose a new formulation of LLM unlearning as a constrained optimization problem: forgetting is enforced via a novel logit-margin flattening loss that explicitly drives the output distribution toward uniformity on a designated forget set, while retention is preserved through a hard constraint on a separate retain set. Compared to entropy-based objectives, our loss is softmax-free, numerically stable, and maintains non-vanishing gradients, enabling more efficient and robust optimization. We solve the constrained problem using a scalable primal-dual algorithm that exposes the trade-off between forgetting and retention through the dynamics of the dual variable. Evaluations on the TOFU and MUSE benchmarks across diverse LLM architectures demonstrate that our approach consistently matches or exceeds state-of-the-art baselines, effectively removing targeted information while preserving downstream utility. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:55:23 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Entesari",
"Taha",
""
],
[
"Hatami",
"Arman",
""
],
[
"Khaziev",
"Rinat",
""
],
[
"Ramakrishna",
"Anil",
""
],
[
"Fazlyab",
"Mahyar",
""
]
] |
2506.05315 | Jeremy Kline | Jeremy B. Kline, Alec Yen, Stanley Chen, Kevin P. O'Brien | The Arm Qubit: A Superconducting Qubit Co-Designed for Coherence and Coupling | 14 pages, 6 figures | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a superconducting qubit which consists of two strongly coupled modes: one for data storage and one for coupling, allowing faster, higher-fidelity entangling gates and readout. The use of a dedicated coupling mode allows nonlinear couplings of several hundred MHz between the data mode and other elements, with minimal linear coupling to the data mode. Including decoherence, simulations show that this architecture enables microwave-only CZ gates with an infidelity of $8.6\times10^{-5}$ in 17 ns and always-on ZZ interaction less than 0.4 kHz. Numerical simulations also show readout with state assignment error of $1\times10^{-4}$ in 27 ns (assuming quantum efficiency $\eta=0.5$), Purcell-limited lifetime of 167 ms without a Purcell filter, and a mechanism to suppress shot-noise dephasing ($1/\Gamma_{\phi}=15.8$ ms). Single-qubit gate infidelities are below $1\times10^{-5}$ including decoherence. These beyond experimental state-of-the-art gate and readout fidelities rely only on capacitive coupling between arm qubits, making the arm qubit a promising scalable building block for fault-tolerant quantum computers. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:55:24 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Kline",
"Jeremy B.",
""
],
[
"Yen",
"Alec",
""
],
[
"Chen",
"Stanley",
""
],
[
"O'Brien",
"Kevin P.",
""
]
] |
2506.05316 | Yifan Sun | Yifan Sun, Jingyan Shen, Yibin Wang, Tianyu Chen, Zhendong Wang, Mingyuan Zhou, Huan Zhang | Improving Data Efficiency for LLM Reinforcement Fine-tuning Through Difficulty-targeted Online Data Selection and Rollout Replay | null | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Reinforcement learning (RL) has become an effective approach for fine-tuning large language models (LLMs), particularly to enhance their reasoning capabilities. However, RL fine-tuning remains highly resource-intensive, and existing work has largely overlooked the problem of data efficiency. In this paper, we propose two techniques to improve data efficiency in LLM RL fine-tuning: difficulty-targeted online data selection and rollout replay. We introduce the notion of adaptive difficulty to guide online data selection, prioritizing questions of moderate difficulty that are more likely to yield informative learning signals. To estimate adaptive difficulty efficiently, we develop an attention-based framework that requires rollouts for only a small reference set of questions. The adaptive difficulty of the remaining questions is then estimated based on their similarity to this set. To further reduce rollout cost, we introduce a rollout replay mechanism that reuses recent rollouts, lowering per-step computation while maintaining stable updates. Extensive experiments across 6 LLM-dataset combinations show that our method reduces RL fine-tuning time by 25% to 65% to reach the same level of performance as the original GRPO algorithm. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:55:43 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Sun",
"Yifan",
""
],
[
"Shen",
"Jingyan",
""
],
[
"Wang",
"Yibin",
""
],
[
"Chen",
"Tianyu",
""
],
[
"Wang",
"Zhendong",
""
],
[
"Zhou",
"Mingyuan",
""
],
[
"Zhang",
"Huan",
""
]
] |
2506.05317 | Daniel Rho | Daniel Rho and Jun Myeong Choi and Biswadip Dey and Roni Sengupta | ProJo4D: Progressive Joint Optimization for Sparse-View Inverse Physics Estimation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural rendering has made significant strides in 3D reconstruction and novel view synthesis. With the integration with physics, it opens up new applications. The inverse problem of estimating physics from visual data, however, still remains challenging, limiting its effectiveness for applications like physically accurate digital twin creation in robotics and XR. Existing methods that incorporate physics into neural rendering frameworks typically require dense multi-view videos as input, making them impractical for scalable, real-world use. When presented with sparse multi-view videos, the sequential optimization strategy used by existing approaches introduces significant error accumulation, e.g., poor initial 3D reconstruction leads to bad material parameter estimation in subsequent stages. Instead of sequential optimization, directly optimizing all parameters at the same time also fails due to the highly non-convex and often non-differentiable nature of the problem. We propose ProJo4D, a progressive joint optimization framework that gradually increases the set of jointly optimized parameters guided by their sensitivity, leading to fully joint optimization over geometry, appearance, physical state, and material property. Evaluations on PAC-NeRF and Spring-Gaus datasets show that ProJo4D outperforms prior work in 4D future state prediction, novel view rendering of future state, and material parameter estimation, demonstrating its effectiveness in physically grounded 4D scene understanding. For demos, please visit the project webpage: https://daniel03c1.github.io/ProJo4D/ | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:55:56 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Rho",
"Daniel",
""
],
[
"Choi",
"Jun Myeong",
""
],
[
"Dey",
"Biswadip",
""
],
[
"Sengupta",
"Roni",
""
]
] |
2506.05318 | Haoyuan Li | Haoyuan Li, Yanpeng Zhou, Yufei Gao, Tao Tang, Jianhua Han, Yujie Yuan, Dave Zhenyu Chen, Jiawang Bian, Hang Xu, Xiaodan Liang | Does Your 3D Encoder Really Work? When Pretrain-SFT from 2D VLMs Meets 3D VLMs | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Remarkable progress in 2D Vision-Language Models (VLMs) has spurred interest in extending them to 3D settings for tasks like 3D Question Answering, Dense Captioning, and Visual Grounding. Unlike 2D VLMs that typically process images through an image encoder, 3D scenes, with their intricate spatial structures, allow for diverse model architectures. Based on their encoder design, this paper categorizes recent 3D VLMs into 3D object-centric, 2D image-based, and 3D scene-centric approaches. Despite the architectural similarity of 3D scene-centric VLMs to their 2D counterparts, they have exhibited comparatively lower performance compared with the latest 3D object-centric and 2D image-based approaches. To understand this gap, we conduct an in-depth analysis, revealing that 3D scene-centric VLMs show limited reliance on the 3D scene encoder, and the pre-train stage appears less effective than in 2D VLMs. Furthermore, we observe that data scaling benefits are less pronounced on larger datasets. Our investigation suggests that while these models possess cross-modal alignment capabilities, they tend to over-rely on linguistic cues and overfit to frequent answer distributions, thereby diminishing the effective utilization of the 3D encoder. To address these limitations and encourage genuine 3D scene understanding, we introduce a novel 3D Relevance Discrimination QA dataset designed to disrupt shortcut learning and improve 3D understanding. Our findings highlight the need for advanced evaluation and improved strategies for better 3D understanding in 3D VLMs. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:56:12 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Li",
"Haoyuan",
""
],
[
"Zhou",
"Yanpeng",
""
],
[
"Gao",
"Yufei",
""
],
[
"Tang",
"Tao",
""
],
[
"Han",
"Jianhua",
""
],
[
"Yuan",
"Yujie",
""
],
[
"Chen",
"Dave Zhenyu",
""
],
[
"Bian",
"Jiawang",
""
],
[
"Xu",
"Hang",
""
],
[
"Liang",
"Xiaodan",
""
]
] |
2506.05319 | Yidun Wan | Yu Zhao and Yidun Wan | Landau-Ginzburg Paradigm of Topological Phases | 50 + 22 pages | null | null | null | cond-mat.str-el hep-th math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Topologically ordered matter phases have been regarded as beyond the Landau-Ginzburg symmetry breaking paradigm of matter phases. Recent studies of anyon condensation in topological phases, however, may fit topological phases back in the Landau-Ginzburg paradigm. To truly do so, we realized that the string-net model of topological phases is in fact an effective lattice gauge theory coupled with anyonic matter once two modifications are made: (1) We reinterpret anyons as matter fields coupled to lattice gauge fields, thus extending the HGW model to a genuine Hamiltonian lattice gauge theory. (2) By explicitly incorporating the internal degrees of freedom of anyons, we construct an enlarged Hilbert space that supports well-defined gauge transformations and covariant coupling, restoring the analogy with conventional lattice gauge field theory. In this modified string-net model, topological phase transitions induced by anyon condensation and their consequent phenomena, such as order parameter fields, coherent states, Goldstone modes, and gapping gauge degrees of freedom, can be formulated exactly as Landau's effective theory of the Higgs mechanism. To facilitate the understanding, we also compare anyon condensation to/with the Higgs boson condensation in the electroweak theory and the Cooper pair condensation. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:57:03 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Zhao",
"Yu",
""
],
[
"Wan",
"Yidun",
""
]
] |
2506.05320 | Nanda H Krishna | Avery Hee-Woon Ryoo, Nanda H. Krishna, Ximeng Mao, Mehdi Azabou, Eva L. Dyer, Matthew G. Perich, Guillaume Lajoie | Generalizable, real-time neural decoding with hybrid state-space models | Preprint. Under review | null | null | null | q-bio.NC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Real-time decoding of neural activity is central to neuroscience and neurotechnology applications, from closed-loop experiments to brain-computer interfaces, where models are subject to strict latency constraints. Traditional methods, including simple recurrent neural networks, are fast and lightweight but often struggle to generalize to unseen data. In contrast, recent Transformer-based approaches leverage large-scale pretraining for strong generalization performance, but typically have much larger computational requirements and are not always suitable for low-resource or real-time settings. To address these shortcomings, we present POSSM, a novel hybrid architecture that combines individual spike tokenization via a cross-attention module with a recurrent state-space model (SSM) backbone to enable (1) fast and causal online prediction on neural activity and (2) efficient generalization to new sessions, individuals, and tasks through multi-dataset pretraining. We evaluate POSSM's decoding performance and inference speed on intracortical decoding of monkey motor tasks, and show that it extends to clinical applications, namely handwriting and speech decoding in human subjects. Notably, we demonstrate that pretraining on monkey motor-cortical recordings improves decoding performance on the human handwriting task, highlighting the exciting potential for cross-species transfer. In all of these tasks, we find that POSSM achieves decoding accuracy comparable to state-of-the-art Transformers, at a fraction of the inference cost (up to 9x faster on GPU). These results suggest that hybrid SSMs are a promising approach to bridging the gap between accuracy, inference speed, and generalization when training neural decoders for real-time, closed-loop applications. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:57:08 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ryoo",
"Avery Hee-Woon",
""
],
[
"Krishna",
"Nanda H.",
""
],
[
"Mao",
"Ximeng",
""
],
[
"Azabou",
"Mehdi",
""
],
[
"Dyer",
"Eva L.",
""
],
[
"Perich",
"Matthew G.",
""
],
[
"Lajoie",
"Guillaume",
""
]
] |
2506.05321 | Maxwell Xu | Maxwell A. Xu, Girish Narayanswamy, Kumar Ayush, Dimitris Spathis, Shun Liao, Shyam A. Tailor, Ahmed Metwally, A. Ali Heydari, Yuwei Zhang, Jake Garrison, Samy Abdel-Ghaffar, Xuhai Xu, Ken Gu, Jacob Sunshine, Ming-Zher Poh, Yun Liu, Tim Althoff, Shrikanth Narayanan, Pushmeet Kohli, Mark Malhotra, Shwetak Patel, Yuzhe Yang, James M. Rehg, Xin Liu, Daniel McDuff | LSM-2: Learning from Incomplete Wearable Sensor Data | Xu and Narayanswamy are co-first authors. McDuff and Liu are co-last authors | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Foundation models, a cornerstone of recent advancements in machine learning, have predominantly thrived on complete and well-structured data. Wearable sensor data frequently suffers from significant missingness, posing a substantial challenge for self-supervised learning (SSL) models that typically assume complete data inputs. This paper introduces the second generation of Large Sensor Model (LSM-2) with Adaptive and Inherited Masking (AIM), a novel SSL approach that learns robust representations directly from incomplete data without requiring explicit imputation. AIM's core novelty lies in its use of learnable mask tokens to model both existing ("inherited") and artificially introduced missingness, enabling it to robustly handle fragmented real-world data during inference. Pre-trained on an extensive dataset of 40M hours of day-long multimodal sensor data, our LSM-2 with AIM achieves the best performance across a diverse range of tasks, including classification, regression and generative modeling. Furthermore, LSM-2 with AIM exhibits superior scaling performance, and critically, maintains high performance even under targeted missingness scenarios, reflecting clinically coherent patterns, such as the diagnostic value of nighttime biosignals for hypertension prediction. This makes AIM a more reliable choice for real-world wearable data applications. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:57:11 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Xu",
"Maxwell A.",
""
],
[
"Narayanswamy",
"Girish",
""
],
[
"Ayush",
"Kumar",
""
],
[
"Spathis",
"Dimitris",
""
],
[
"Liao",
"Shun",
""
],
[
"Tailor",
"Shyam A.",
""
],
[
"Metwally",
"Ahmed",
""
],
[
"Heydari",
"A. Ali",
""
],
[
"Zhang",
"Yuwei",
""
],
[
"Garrison",
"Jake",
""
],
[
"Abdel-Ghaffar",
"Samy",
""
],
[
"Xu",
"Xuhai",
""
],
[
"Gu",
"Ken",
""
],
[
"Sunshine",
"Jacob",
""
],
[
"Poh",
"Ming-Zher",
""
],
[
"Liu",
"Yun",
""
],
[
"Althoff",
"Tim",
""
],
[
"Narayanan",
"Shrikanth",
""
],
[
"Kohli",
"Pushmeet",
""
],
[
"Malhotra",
"Mark",
""
],
[
"Patel",
"Shwetak",
""
],
[
"Yang",
"Yuzhe",
""
],
[
"Rehg",
"James M.",
""
],
[
"Liu",
"Xin",
""
],
[
"McDuff",
"Daniel",
""
]
] |
2506.05322 | Alexandros Hollender | Aris Filos-Ratsikas, Yiannis Giannakopoulos, Alexandros Hollender, Charalampos Kokkalis | Equilibrium Computation in First-Price Auctions with Correlated Priors | null | null | null | null | cs.GT cs.CC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the computational complexity of computing Bayes-Nash equilibria in first-price auctions, where the bidders' values for the item are drawn from a general (possibly correlated) joint distribution. We show that when the values and the bidding space are discrete, determining the existence of a pure Bayes-Nash equilibrium is NP-hard. This is the first hardness result in the literature of the problem that does not rely on assumptions of subjectivity of the priors, or convoluted tie-breaking rules. We then present two main approaches for achieving positive results, via bid sparsification and via bid densification. The former is more combinatorial and is based on enumeration techniques, whereas the latter makes use of the continuous theory of the problem developed in the economics literature. Using these approaches, we develop polynomial-time approximation algorithms for computing equilibria in symmetric settings or settings with a fixed number of bidders, for different (discrete or continuous) variants of the auction. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:57:16 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Filos-Ratsikas",
"Aris",
""
],
[
"Giannakopoulos",
"Yiannis",
""
],
[
"Hollender",
"Alexandros",
""
],
[
"Kokkalis",
"Charalampos",
""
]
] |
2506.05323 | David Headley | David Headley, Nicholas Chancellor | Non-Perturbative Topological Gadgets for Many-Body Coupling | 14 pages, 8 figures | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Continuous-time quantum hardware implementations generally lack the native capability to implement high-order terms that would facilitate efficient compilation of quantum algorithms. This limitation has, in part, motivated the development of perturbative gadgets -- multi-qubit constructions used to effect a desired Hamiltonian using engineered low-energy subspaces of a larger system constructed using simpler, usually two-body, primitives. In this work, we demonstrate how a class of non-perturbative gadgets can produce high-order multi-body interactions by taking advantage of the odd-even properties of topological defect subspaces. The simplest example is based on domain-wall defects forming an effective Ising spin-chain based on three-body coupling with linear connectivity, alongside three-, or five-body driving terms depending on the intended use. Although this will be the main focus of the paper due to conceptual simplicity, there exist systems constructed with only two-body couplings where the boundaries determine whether there are an odd or even number of defects, namely ice-like systems which may yield more complex gadget-like constructions. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:57:32 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Headley",
"David",
""
],
[
"Chancellor",
"Nicholas",
""
]
] |
2506.05324 | Ling Yan Hung Dr | Ling-Yan Hung, Kaixin Ji, Ce Shen, Yidun Wan and Yu Zhao | A 2D-CFT Factory: Critical Lattice Models from Competing Anyon Condensation Processes in SymTO/SymTFT | 47 pages + appendices, 20 figures | null | null | null | cond-mat.str-el cond-mat.stat-mech hep-th math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduce a ``CFT factory'' : a novel algorithm of methodically generating 2D lattice models that would flow to 2D conformal fixed points in the infrared. These 2D models are realised by giving critical boundary conditions to 3D topological orders (symTOs/symTFTs) described by string-net models, often called the strange correlators. We engineer these critical boundary conditions by introducing a commensurate amount of non-commuting anyon condensates. The non-invertible symmetries preserved at the critical point can be controlled by studying a novel ``refined condensation tree''. Our structured method generates an infinite family of critical lattice models, including the A-series minimal models, and uncovers previously unknown critical points. Notably, we find at least three novel critical points (c$\approx 1.3$, $1.8$, and $2.5$ respectively) preserving the Haagerup symmetries, in addition to recovering previously reported ones. The condensation tree, together with a generalised Kramers-Wannier duality, predicts precisely large swathes of phase boundaries, fixes almost completely the global phase diagram, and sieves out second order phase transitions. This is not only illustrated in well-known examples (such as the 8-vertex model related to the $A_5$ category) but also further verified with precision numerics, using our improved (non-invertible) symmetry-preserving tensor-network RG, in novel examples involving the Haagerup symmetries. We show that critical couplings can be precisely encoded in the categorical data (Frobenius algebras and quantum dimensions in unitary fusion categories), thus establishing a powerful, systematic route to discovering and potentially classifying new conformal field theories. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:04 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Hung",
"Ling-Yan",
""
],
[
"Ji",
"Kaixin",
""
],
[
"Shen",
"Ce",
""
],
[
"Wan",
"Yidun",
""
],
[
"Zhao",
"Yu",
""
]
] |
2506.05325 | Yingshuai Ji | Yingshuai Ji, Haomin Zhuang, Matthew Toole, James McKenzie, Xiaolong Liu and Xiangliang Zhang | Seeing the Invisible: Machine learning-Based QPI Kernel Extraction via Latent Alignment | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quasiparticle interference (QPI) imaging is a powerful tool for probing electronic structures in quantum materials, but extracting the single-scatterer QPI pattern (i.e., the kernel) from a multi-scatterer image remains a fundamentally ill-posed inverse problem. In this work, we propose the first AI-based framework for QPI kernel extraction. We introduce a two-step learning strategy that decouples kernel representation learning from observation-to-kernel inference. In the first step, we train a variational autoencoder to learn a compact latent space of scattering kernels. In the second step, we align the latent representation of QPI observations with those of the pre-learned kernels using a dedicated encoder. This design enables the model to infer kernels robustly even under complex, entangled scattering conditions. We construct a diverse and physically realistic QPI dataset comprising 100 unique kernels and evaluate our method against a direct one-step baseline. Experimental results demonstrate that our approach achieves significantly higher extraction accuracy, and improved generalization to unseen kernels. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:09 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ji",
"Yingshuai",
""
],
[
"Zhuang",
"Haomin",
""
],
[
"Toole",
"Matthew",
""
],
[
"McKenzie",
"James",
""
],
[
"Liu",
"Xiaolong",
""
],
[
"Zhang",
"Xiangliang",
""
]
] |
2506.05326 | Ram Purandhar Reddy Sudha | Sagar Airen, Zackaria Chacko, Can Kilic, Ram Purandhar Reddy Sudha | Searching for Hidden Sector Particles at Neutrino Telescopes | 29 pages, 18 figures and 1 table | null | null | UT-WI-41-2024 | hep-ph hep-ex | http://creativecommons.org/licenses/by/4.0/ | We explore the possibility of directly detecting light, long-lived hidden sector particles at the IceCube neutrino telescope. Such particles frequently arise in non-minimal hidden sectors that couple to the Standard Model through portal operators. We consider two distinct scenarios. In the first scenario, which arises from a neutrino portal interaction, a hidden sector particle is produced inside the detector by the collision of an energetic neutrino with a nucleon, giving rise to a visible cascade. This new state then decays into a hidden sector daughter, which can naturally be long-lived. The eventual decay of the daughter particle back to Standard Model states gives rise to a second cascade inside the detector. This scenario therefore gives rise to a characteristic "double bang" signal arising from the two distinct cascades. In the second scenario, which arises from a hypercharge portal interaction, a hidden sector particle is produced outside the detector by the collision of an atmospheric muon with a nucleon. This new state promptly decays into a pair of hidden sector daughters that are long-lived. If both daughters decay into Standard Model states inside the detector, we again obtain a double-bang signal from the two distinct cascades. We explore the reach of IceCube for these two scenarios and show that it has the potential to significantly improve the sensitivity to hidden sector models in the mass range from about a GeV to about 20 GeV. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:15 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Airen",
"Sagar",
""
],
[
"Chacko",
"Zackaria",
""
],
[
"Kilic",
"Can",
""
],
[
"Sudha",
"Ram Purandhar Reddy",
""
]
] |
2506.05327 | Duochao Shi | Duochao Shi, Weijie Wang, Donny Y. Chen, Zeyu Zhang, Jia-Wang Bian, Bohan Zhuang, Chunhua Shen | Revisiting Depth Representations for Feed-Forward 3D Gaussian Splatting | Project page: https://aim-uofa.github.io/PMLoss | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Depth maps are widely used in feed-forward 3D Gaussian Splatting (3DGS) pipelines by unprojecting them into 3D point clouds for novel view synthesis. This approach offers advantages such as efficient training, the use of known camera poses, and accurate geometry estimation. However, depth discontinuities at object boundaries often lead to fragmented or sparse point clouds, degrading rendering quality -- a well-known limitation of depth-based representations. To tackle this issue, we introduce PM-Loss, a novel regularization loss based on a pointmap predicted by a pre-trained transformer. Although the pointmap itself may be less accurate than the depth map, it effectively enforces geometric smoothness, especially around object boundaries. With the improved depth map, our method significantly improves the feed-forward 3DGS across various architectures and scenes, delivering consistently better rendering results. Our project page: https://aim-uofa.github.io/PMLoss | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:23 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Shi",
"Duochao",
""
],
[
"Wang",
"Weijie",
""
],
[
"Chen",
"Donny Y.",
""
],
[
"Zhang",
"Zeyu",
""
],
[
"Bian",
"Jia-Wang",
""
],
[
"Zhuang",
"Bohan",
""
],
[
"Shen",
"Chunhua",
""
]
] |
2506.05328 | Lidong Lu | Lidong Lu, Guo Chen, Zhiqi Li, Yicheng Liu, Tong Lu | AV-Reasoner: Improving and Benchmarking Clue-Grounded Audio-Visual Counting for MLLMs | 21 pages, 11 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite progress in video understanding, current MLLMs struggle with counting tasks. Existing benchmarks are limited by short videos, close-set queries, lack of clue annotations, and weak multimodal coverage. In this paper, we introduce CG-AV-Counting, a manually-annotated clue-grounded counting benchmark with 1,027 multimodal questions and 5,845 annotated clues over 497 long videos. It supports both black-box and white-box evaluation, serving as a comprehensive testbed for both end-to-end and reasoning-based counting. To explore ways to improve model's counting capability, we propose AV-Reasoner, a model trained with GRPO and curriculum learning to generalize counting ability from related tasks. AV-Reasoner achieves state-of-the-art results across multiple benchmarks, demonstrating the effectiveness of reinforcement learning. However, experiments show that on out-of-domain benchmarks, reasoning in the language space fails to bring performance gains. The code and benchmark have been realeased on https://av-reasoner.github.io. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:33 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Lu",
"Lidong",
""
],
[
"Chen",
"Guo",
""
],
[
"Li",
"Zhiqi",
""
],
[
"Liu",
"Yicheng",
""
],
[
"Lu",
"Tong",
""
]
] |
2506.05329 | Chao Qin | Guido Imbens, Chao Qin, Stefan Wager | Admissibility of Completely Randomized Trials: A Large-Deviation Approach | A one-page abstract of this work will appear at the 26th ACM Conference on Economics and Computation (EC'25) | null | null | null | stat.ML cs.LG econ.EM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When an experimenter has the option of running an adaptive trial, is it admissible to ignore this option and run a non-adaptive trial instead? We provide a negative answer to this question in the best-arm identification problem, where the experimenter aims to allocate measurement efforts judiciously to confidently deploy the most effective treatment arm. We find that, whenever there are at least three treatment arms, there exist simple adaptive designs that universally and strictly dominate non-adaptive completely randomized trials. This dominance is characterized by a notion called efficiency exponent, which quantifies a design's statistical efficiency when the experimental sample is large. Our analysis focuses on the class of batched arm elimination designs, which progressively eliminate underperforming arms at pre-specified batch intervals. We characterize simple sufficient conditions under which these designs universally and strictly dominate completely randomized trials. These results resolve the second open problem posed in Qin [2022]. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:43 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Imbens",
"Guido",
""
],
[
"Qin",
"Chao",
""
],
[
"Wager",
"Stefan",
""
]
] |
2506.05330 | Miguel Gon\c{c}alves | Miguel Gon\c{c}alves, Juan Felipe Mendez-Valderrama, Jonah Herzog-Arbeitman, Jiabin Yu, Xiaodong Xu, Di Xiao, B. Andrei Bernevig, Nicolas Regnault | Spinless and spinful charge excitations in moir\'e Fractional Chern Insulators | null | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | Fractionally charged elementary excitations, the quasi-electron and quasi-hole, are one of the hallmarks of the fractional Chern insulator (FCI). In this work, we observe that spontaneous spin polarization in twisted MoTe$_2$ leads to multiple species of low-energy quasi-particles distinguished by their spin quantum numbers. We perform large-scale exact diagonalization (ED) calculations to investigate the nature of these excitations and develop a method to extract their fundamental energetic properties. Focusing on $\theta = 3.7^{\circ}$ and filling factor $\nu = -2/3$ relevant to recent experiments, we show that spin-preserving (spinless) charge excitations have smaller gap than spin-flipping (spinful) excitations both with and without band mixing. This result is in qualitative agreement with the measured magnetic field dependence of the transport gaps. Beyond the spinless and spinful quasi-particle gaps, we extract the full quasi-electron and quasi-hole ``band structure'' and find significant dispersion with emergent magnetic translation symmetry -- a fundamental departure from the immobile excitations of the quantum Hall fluid. Our results establish a framework for computing the properties of novel elementary excitations in FCIs. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:58:45 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Gonçalves",
"Miguel",
""
],
[
"Mendez-Valderrama",
"Juan Felipe",
""
],
[
"Herzog-Arbeitman",
"Jonah",
""
],
[
"Yu",
"Jiabin",
""
],
[
"Xu",
"Xiaodong",
""
],
[
"Xiao",
"Di",
""
],
[
"Bernevig",
"B. Andrei",
""
],
[
"Regnault",
"Nicolas",
""
]
] |
2506.05331 | Xinyan Chen | Xinyan Chen, Renrui Zhang, Dongzhi Jiang, Aojun Zhou, Shilin Yan, Weifeng Lin, Hongsheng Li | MINT-CoT: Enabling Interleaved Visual Tokens in Mathematical Chain-of-Thought Reasoning | Code is released at https://github.com/xinyan-cxy/MINT-CoT | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Chain-of-Thought (CoT) has widely enhanced mathematical reasoning in Large Language Models (LLMs), but it still remains challenging for extending it to multimodal domains. Existing works either adopt a similar textual reasoning for image input, or seek to interleave visual signals into mathematical CoT. However, they face three key limitations for math problem-solving: reliance on coarse-grained box-shaped image regions, limited perception of vision encoders on math content, and dependence on external capabilities for visual modification. In this paper, we propose MINT-CoT, introducing Mathematical INterleaved Tokens for Chain-of-Thought visual reasoning. MINT-CoT adaptively interleaves relevant visual tokens into textual reasoning steps via an Interleave Token, which dynamically selects visual regions of any shapes within math figures. To empower this capability, we construct the MINT-CoT dataset, containing 54K mathematical problems aligning each reasoning step with visual regions at the token level, accompanied by a rigorous data generation pipeline. We further present a three-stage MINT-CoT training strategy, progressively combining text-only CoT SFT, interleaved CoT SFT, and interleaved CoT RL, which derives our MINT-CoT-7B model. Extensive experiments demonstrate the effectiveness of our method for effective visual interleaved reasoning in mathematical domains, where MINT-CoT-7B outperforms the baseline model by +34.08% on MathVista, +28.78% on GeoQA, and +23.2% on MMStar, respectively. Our code and data are available at https://github.com/xinyan-cxy/MINT-CoT | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:02 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Chen",
"Xinyan",
""
],
[
"Zhang",
"Renrui",
""
],
[
"Jiang",
"Dongzhi",
""
],
[
"Zhou",
"Aojun",
""
],
[
"Yan",
"Shilin",
""
],
[
"Lin",
"Weifeng",
""
],
[
"Li",
"Hongsheng",
""
]
] |
2506.05332 | Jingyang Lin | Jingyang Lin, Jialian Wu, Ximeng Sun, Ze Wang, Jiang Liu, Yusheng Su, Xiaodong Yu, Hao Chen, Jiebo Luo, Zicheng Liu, Emad Barsoum | Unleashing Hour-Scale Video Training for Long Video-Language Understanding | Project page: https://videomarathon.github.io/ | null | null | null | cs.CV cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent long-form video-language understanding benchmarks have driven progress in video large multimodal models (Video-LMMs). However, the scarcity of well-annotated long videos has left the training of hour-long Video-LLMs underexplored. To close this gap, we present VideoMarathon, a large-scale hour-long video instruction-following dataset. This dataset includes around 9,700 hours of long videos sourced from diverse domains, ranging from 3 to 60 minutes per video. Specifically, it contains 3.3M high-quality QA pairs, spanning six fundamental topics: temporality, spatiality, object, action, scene, and event. Compared to existing video instruction datasets, VideoMarathon significantly extends training video durations up to 1 hour, and supports 22 diverse tasks requiring both short- and long-term video comprehension. Building on VideoMarathon, we propose Hour-LLaVA, a powerful and efficient Video-LMM for hour-scale video-language modeling. It enables hour-long video training and inference at 1-FPS sampling by leveraging a memory augmentation module, which adaptively integrates user question-relevant and spatiotemporal-informative semantics from a cached full video context. In our experiments, Hour-LLaVA achieves the best performance on multiple long video-language benchmarks, demonstrating the high quality of the VideoMarathon dataset and the superiority of the Hour-LLaVA model. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:04 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Lin",
"Jingyang",
""
],
[
"Wu",
"Jialian",
""
],
[
"Sun",
"Ximeng",
""
],
[
"Wang",
"Ze",
""
],
[
"Liu",
"Jiang",
""
],
[
"Su",
"Yusheng",
""
],
[
"Yu",
"Xiaodong",
""
],
[
"Chen",
"Hao",
""
],
[
"Luo",
"Jiebo",
""
],
[
"Liu",
"Zicheng",
""
],
[
"Barsoum",
"Emad",
""
]
] |
2506.05333 | Zhuoming Chen | Ranajoy Sadhukhan, Zhuoming Chen, Haizhong Zheng, Yang Zhou, Emma Strubell, Beidi Chen | Kinetics: Rethinking Test-Time Scaling Laws | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | We rethink test-time scaling laws from a practical efficiency perspective, revealing that the effectiveness of smaller models is significantly overestimated. Prior work, grounded in compute-optimality, overlooks critical memory access bottlenecks introduced by inference-time strategies (e.g., Best-of-$N$, long CoTs). Our holistic analysis, spanning models from 0.6B to 32B parameters, reveals a new Kinetics Scaling Law that better guides resource allocation by incorporating both computation and memory access costs. Kinetics Scaling Law suggests that test-time compute is more effective when used on models above a threshold than smaller ones. A key reason is that in TTS, attention, rather than parameter count, emerges as the dominant cost factor. Motivated by this, we propose a new scaling paradigm centered on sparse attention, which lowers per-token cost and enables longer generations and more parallel samples within the same resource budget. Empirically, we show that sparse attention models consistently outperform dense counterparts, achieving over 60 points gains in low-cost regimes and over 5 points gains in high-cost regimes for problem-solving accuracy on AIME, encompassing evaluations on state-of-the-art MoEs. These results suggest that sparse attention is essential for realizing the full potential of test-time scaling because, unlike training, where parameter scaling saturates, test-time accuracy continues to improve through increased generation. The code is available at https://github.com/Infini-AI-Lab/Kinetics. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:24 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Sadhukhan",
"Ranajoy",
""
],
[
"Chen",
"Zhuoming",
""
],
[
"Zheng",
"Haizhong",
""
],
[
"Zhou",
"Yang",
""
],
[
"Strubell",
"Emma",
""
],
[
"Chen",
"Beidi",
""
]
] |
2506.05334 | Tsung-Han Wu | Mihran Miroyan, Tsung-Han Wu, Logan King, Tianle Li, Jiayi Pan, Xinyan Hu, Wei-Lin Chiang, Anastasios N. Angelopoulos, Trevor Darrell, Narges Norouzi, Joseph E. Gonzalez | Search Arena: Analyzing Search-Augmented LLMs | Preprint. Code: https://github.com/lmarena/search-arena. Dataset: https://huggingface.co/datasets/lmarena-ai/search-arena-24k | null | null | null | cs.CL cs.IR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Search-augmented language models combine web search with Large Language Models (LLMs) to improve response groundedness and freshness. However, analyzing these systems remains challenging: existing datasets are limited in scale and narrow in scope, often constrained to static, single-turn, fact-checking questions. In this work, we introduce Search Arena, a crowd-sourced, large-scale, human-preference dataset of over 24,000 paired multi-turn user interactions with search-augmented LLMs. The dataset spans diverse intents and languages, and contains full system traces with around 12,000 human preference votes. Our analysis reveals that user preferences are influenced by the number of citations, even when the cited content does not directly support the attributed claims, uncovering a gap between perceived and actual credibility. Furthermore, user preferences vary across cited sources, revealing that community-driven platforms are generally preferred and static encyclopedic sources are not always appropriate and reliable. To assess performance across different settings, we conduct cross-arena analyses by testing search-augmented LLMs in a general-purpose chat environment and conventional LLMs in search-intensive settings. We find that web search does not degrade and may even improve performance in non-search settings; however, the quality in search settings is significantly affected if solely relying on the model's parametric knowledge. We open-sourced the dataset to support future research in this direction. Our dataset and code are available at: https://github.com/lmarena/search-arena. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:26 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Miroyan",
"Mihran",
""
],
[
"Wu",
"Tsung-Han",
""
],
[
"King",
"Logan",
""
],
[
"Li",
"Tianle",
""
],
[
"Pan",
"Jiayi",
""
],
[
"Hu",
"Xinyan",
""
],
[
"Chiang",
"Wei-Lin",
""
],
[
"Angelopoulos",
"Anastasios N.",
""
],
[
"Darrell",
"Trevor",
""
],
[
"Norouzi",
"Narges",
""
],
[
"Gonzalez",
"Joseph E.",
""
]
] |
2506.05335 | Maxim Shirokov Evgenyevich | M.E.Shirokov | Upper bound for the Holevo quantity arising from the fundamental entropic inequality | 7 pages, any comments and reference are welcome | null | null | null | quant-ph cs.IT math-ph math.IT math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show how the fundamental entropic inequality proved recently in [arXiv:2408.15306] can be used to obtain a quite accurate upper bound on the Holevo quantity of a discrete ensemble of quantum states expressed via the probabilities and the metric characteristics of this ensembles. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:26 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Shirokov",
"M. E.",
""
]
] |
2506.05336 | Mohammad Hanan Gani | Ghazi Shazan Ahmad, Ahmed Heakl, Hanan Gani, Abdelrahman Shaker, Zhiqiang Shen, Ranjay Krishna, Fahad Shahbaz Khan and Salman Khan | VideoMolmo: Spatio-Temporal Grounding Meets Pointing | 20 pages, 13 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Spatio-temporal localization is vital for precise interactions across diverse domains, from biological research to autonomous navigation and interactive interfaces. Current video-based approaches, while proficient in tracking, lack the sophisticated reasoning capabilities of large language models, limiting their contextual understanding and generalization. We introduce VideoMolmo, a large multimodal model tailored for fine-grained spatio-temporal pointing conditioned on textual descriptions. Building upon the Molmo architecture, VideoMolmo incorporates a temporal module utilizing an attention mechanism to condition each frame on preceding frames, ensuring temporal consistency. Additionally, our novel temporal mask fusion pipeline employs SAM2 for bidirectional point propagation, significantly enhancing coherence across video sequences. This two-step decomposition, i.e., first using the LLM to generate precise pointing coordinates, then relying on a sequential mask-fusion module to produce coherent segmentation, not only simplifies the task for the language model but also enhances interpretability. Due to the lack of suitable datasets, we curate a comprehensive dataset comprising 72k video-caption pairs annotated with 100k object points. To evaluate the generalization of VideoMolmo, we introduce VPoS-Bench, a challenging out-of-distribution benchmark spanning five real-world scenarios: Cell Tracking, Egocentric Vision, Autonomous Driving, Video-GUI Interaction, and Robotics. We also evaluate our model on Referring Video Object Segmentation (Refer-VOS) and Reasoning VOS tasks. In comparison to existing models, VideoMolmo substantially improves spatio-temporal pointing accuracy and reasoning capability. Our code and models are publicly available at https://github.com/mbzuai-oryx/VideoMolmo. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:29 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ahmad",
"Ghazi Shazan",
""
],
[
"Heakl",
"Ahmed",
""
],
[
"Gani",
"Hanan",
""
],
[
"Shaker",
"Abdelrahman",
""
],
[
"Shen",
"Zhiqiang",
""
],
[
"Krishna",
"Ranjay",
""
],
[
"Khan",
"Fahad Shahbaz",
""
],
[
"Khan",
"Salman",
""
]
] |
2506.05337 | Emmanuel de Oliveira | Joao Vitor C. Lovato, Edgar Huayra, Emmanuel G. de Oliveira | Momentum fraction and hard scale dependence of double parton scattering | 11 pages, 4 figures | null | null | null | hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The effective cross section of double parton scattering in high-energy hadron collisions has been measured in proton--proton collisions, with significant variation among final-state observables, contrary to the idea of a universal value. Building upon our previous work, we incorporate the dependence on both the parton longitudinal momentum fraction $x$ and the process energy hard scale $\mu$ into the transverse part of the double parton distributions, using a Gaussian profile. Employing the experimental data from the LHC and Tevatron experiments (covering different processes, kinematic configurations, and center--of--mass energies), we perform a global fit of the model, extracting the parameters that describe the proton structure. With this result, it becomes possible to calculate the effective cross section for others observables, and we provide predictions for future measurements at the LHC. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:30 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Lovato",
"Joao Vitor C.",
""
],
[
"Huayra",
"Edgar",
""
],
[
"de Oliveira",
"Emmanuel G.",
""
]
] |
2506.05338 | Alan Dolhasz | Alan Dolhasz, Chen Ma, Dave Gausebeck, Kevin Chen, Gregor Miller, Lucas Hayne, Gunnar Hovden, Azwad Sabik, Olaf Brandt, Mira Slavcheva | Defurnishing with X-Ray Vision: Joint Removal of Furniture from Panoramas and Mesh | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present a pipeline for generating defurnished replicas of indoor spaces represented as textured meshes and corresponding multi-view panoramic images. To achieve this, we first segment and remove furniture from the mesh representation, extend planes, and fill holes, obtaining a simplified defurnished mesh (SDM). This SDM acts as an ``X-ray'' of the scene's underlying structure, guiding the defurnishing process. We extract Canny edges from depth and normal images rendered from the SDM. We then use these as a guide to remove the furniture from panorama images via ControlNet inpainting. This control signal ensures the availability of global geometric information that may be hidden from a particular panoramic view by the furniture being removed. The inpainted panoramas are used to texture the mesh. We show that our approach produces higher quality assets than methods that rely on neural radiance fields, which tend to produce blurry low-resolution images, or RGB-D inpainting, which is highly susceptible to hallucinations. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:30 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Dolhasz",
"Alan",
""
],
[
"Ma",
"Chen",
""
],
[
"Gausebeck",
"Dave",
""
],
[
"Chen",
"Kevin",
""
],
[
"Miller",
"Gregor",
""
],
[
"Hayne",
"Lucas",
""
],
[
"Hovden",
"Gunnar",
""
],
[
"Sabik",
"Azwad",
""
],
[
"Brandt",
"Olaf",
""
],
[
"Slavcheva",
"Mira",
""
]
] |
2506.05339 | Chaitanya Malaviya | Anirudh Bharadwaj, Chaitanya Malaviya, Nitish Joshi, Mark Yatskar | Flattery, Fluff, and Fog: Diagnosing and Mitigating Idiosyncratic Biases in Preference Models | Code and data available at https://github.com/anirudhb123/preference-model-biases | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Language models serve as proxies for human preference judgements in alignment and evaluation, yet they exhibit systematic miscalibration, prioritizing superficial patterns over substantive qualities. This bias manifests as overreliance on features like length, structure, and style, leading to issues like reward hacking and unreliable evaluations. Evidence suggests these biases originate in artifacts in human training data. In this work, we systematically investigate the relationship between training data biases and preference model miscalibration across five idiosyncratic features of language model generations: length, structure, jargon, sycophancy and vagueness. Using controlled counterfactual pairs, we first quantify the extent to which preference models favor responses with magnified biases (skew), finding this preference occurs in >60% of instances, and model preferences show high miscalibration (~40%) compared to human preferences. Notably, bias features only show mild negative correlations to human preference labels (mean r_human = -0.12) but show moderately strong positive correlations with labels from a strong reward model (mean r_model = +0.36), suggesting that models may overrely on spurious cues. To mitigate these issues, we propose a simple post-training method based on counterfactual data augmentation (CDA) using synthesized contrastive examples. Finetuning models with CDA reduces average miscalibration from 39.4% to 32.5% and average absolute skew difference from 20.5% to 10.0%, while maintaining overall RewardBench performance, showing that targeted debiasing is effective for building reliable preference models. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:32 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Bharadwaj",
"Anirudh",
""
],
[
"Malaviya",
"Chaitanya",
""
],
[
"Joshi",
"Nitish",
""
],
[
"Yatskar",
"Mark",
""
]
] |
2506.05340 | Keshigeyan Chandrasegaran | Keshigeyan Chandrasegaran, Michael Poli, Daniel Y. Fu, Dongjun Kim, Lea M. Hadzic, Manling Li, Agrim Gupta, Stefano Massaroli, Azalia Mirhoseini, Juan Carlos Niebles, Stefano Ermon, Li Fei-Fei | Exploring Diffusion Transformer Designs via Grafting | 22 pages; Project website: https://grafting.stanford.edu | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Designing model architectures requires decisions such as selecting operators (e.g., attention, convolution) and configurations (e.g., depth, width). However, evaluating the impact of these decisions on model quality requires costly pretraining, limiting architectural investigation. Inspired by how new software is built on existing code, we ask: can new architecture designs be studied using pretrained models? To this end, we present grafting, a simple approach for editing pretrained diffusion transformers (DiTs) to materialize new architectures under small compute budgets. Informed by our analysis of activation behavior and attention locality, we construct a testbed based on the DiT-XL/2 design to study the impact of grafting on model quality. Using this testbed, we develop a family of hybrid designs via grafting: replacing softmax attention with gated convolution, local attention, and linear attention, and replacing MLPs with variable expansion ratio and convolutional variants. Notably, many hybrid designs achieve good quality (FID: 2.38-2.64 vs. 2.27 for DiT-XL/2) using <2% pretraining compute. We then graft a text-to-image model (PixArt-Sigma), achieving a 1.43x speedup with less than a 2% drop in GenEval score. Finally, we present a case study that restructures DiT-XL/2 by converting every pair of sequential transformer blocks into parallel blocks via grafting. This reduces model depth by 2x and yields better quality (FID: 2.77) than other models of comparable depth. Together, we show that new diffusion model designs can be explored by grafting pretrained DiTs, with edits ranging from operator replacement to architecture restructuring. Code and grafted models: https://grafting.stanford.edu | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:40 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Chandrasegaran",
"Keshigeyan",
""
],
[
"Poli",
"Michael",
""
],
[
"Fu",
"Daniel Y.",
""
],
[
"Kim",
"Dongjun",
""
],
[
"Hadzic",
"Lea M.",
""
],
[
"Li",
"Manling",
""
],
[
"Gupta",
"Agrim",
""
],
[
"Massaroli",
"Stefano",
""
],
[
"Mirhoseini",
"Azalia",
""
],
[
"Niebles",
"Juan Carlos",
""
],
[
"Ermon",
"Stefano",
""
],
[
"Fei-Fei",
"Li",
""
]
] |
2506.05341 | Xingjian Ran | Xingjian Ran, Yixuan Li, Linning Xu, Mulin Yu, Bo Dai | Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning | Project Page: https://directlayout.github.io/ | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Realistic 3D indoor scene synthesis is vital for embodied AI and digital content creation. It can be naturally divided into two subtasks: object generation and layout generation. While recent generative models have significantly advanced object-level quality and controllability, layout generation remains challenging due to limited datasets. Existing methods either overfit to these datasets or rely on predefined constraints to optimize numerical layout that sacrifice flexibility. As a result, they fail to generate scenes that are both open-vocabulary and aligned with fine-grained user instructions. We introduce DirectLayout, a framework that directly generates numerical 3D layouts from text descriptions using generalizable spatial reasoning of large language models (LLMs). DirectLayout decomposes the generation into three stages: producing a Bird's-Eye View (BEV) layout, lifting it into 3D space, and refining object placements. To enable explicit spatial reasoning and help the model grasp basic principles of object placement, we employ Chain-of-Thought (CoT) Activation based on the 3D-Front dataset. Additionally, we design CoT-Grounded Generative Layout Reward to enhance generalization and spatial planning. During inference, DirectLayout addresses asset-layout mismatches via Iterative Asset-Layout Alignment through in-context learning. Extensive experiments demonstrate that DirectLayout achieves impressive semantic consistency, generalization and physical plausibility. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:42 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ran",
"Xingjian",
""
],
[
"Li",
"Yixuan",
""
],
[
"Xu",
"Linning",
""
],
[
"Yu",
"Mulin",
""
],
[
"Dai",
"Bo",
""
]
] |
2506.05342 | Shengcao Cao | Shengcao Cao, Zijun Wei, Jason Kuen, Kangning Liu, Lingzhi Zhang, Jiuxiang Gu, HyunJoon Jung, Liang-Yan Gui, Yu-Xiong Wang | Refer to Anything with Vision-Language Prompts | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Recent image segmentation models have advanced to segment images into high-quality masks for visual entities, and yet they cannot provide comprehensive semantic understanding for complex queries based on both language and vision. This limitation reduces their effectiveness in applications that require user-friendly interactions driven by vision-language prompts. To bridge this gap, we introduce a novel task of omnimodal referring expression segmentation (ORES). In this task, a model produces a group of masks based on arbitrary prompts specified by text only or text plus reference visual entities. To address this new challenge, we propose a novel framework to "Refer to Any Segmentation Mask Group" (RAS), which augments segmentation models with complex multimodal interactions and comprehension via a mask-centric large multimodal model. For training and benchmarking ORES models, we create datasets MaskGroups-2M and MaskGroups-HQ to include diverse mask groups specified by text and reference entities. Through extensive evaluation, we demonstrate superior performance of RAS on our new ORES task, as well as classic referring expression segmentation (RES) and generalized referring expression segmentation (GRES) tasks. Project page: https://Ref2Any.github.io. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:51 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Cao",
"Shengcao",
""
],
[
"Wei",
"Zijun",
""
],
[
"Kuen",
"Jason",
""
],
[
"Liu",
"Kangning",
""
],
[
"Zhang",
"Lingzhi",
""
],
[
"Gu",
"Jiuxiang",
""
],
[
"Jung",
"HyunJoon",
""
],
[
"Gui",
"Liang-Yan",
""
],
[
"Wang",
"Yu-Xiong",
""
]
] |
2506.05343 | Renjie Chen | Wenfeng Lin, Renjie Chen, Boyuan Liu, Shiyue Yan, Ruoyu Feng, Jiangchuan Wei, Yichen Zhang, Yimeng Zhou, Chao Feng, Jiao Ran, Qi Wu, Zuotao Liu, Mingyu Guo | ContentV: Efficient Training of Video Generation Models with Limited Compute | Project Page: https://contentv.github.io | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in video generation demand increasingly efficient training recipes to mitigate escalating computational costs. In this report, we present ContentV, an 8B-parameter text-to-video model that achieves state-of-the-art performance (85.14 on VBench) after training on 256 x 64GB Neural Processing Units (NPUs) for merely four weeks. ContentV generates diverse, high-quality videos across multiple resolutions and durations from text prompts, enabled by three key innovations: (1) A minimalist architecture that maximizes reuse of pre-trained image generation models for video generation; (2) A systematic multi-stage training strategy leveraging flow matching for enhanced efficiency; and (3) A cost-effective reinforcement learning with human feedback framework that improves generation quality without requiring additional human annotations. All the code and models are available at: https://contentv.github.io. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:54 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Lin",
"Wenfeng",
""
],
[
"Chen",
"Renjie",
""
],
[
"Liu",
"Boyuan",
""
],
[
"Yan",
"Shiyue",
""
],
[
"Feng",
"Ruoyu",
""
],
[
"Wei",
"Jiangchuan",
""
],
[
"Zhang",
"Yichen",
""
],
[
"Zhou",
"Yimeng",
""
],
[
"Feng",
"Chao",
""
],
[
"Ran",
"Jiao",
""
],
[
"Wu",
"Qi",
""
],
[
"Liu",
"Zuotao",
""
],
[
"Guo",
"Mingyu",
""
]
] |
2506.05344 | Zuyan Liu | Jiahui Wang, Zuyan Liu, Yongming Rao, Jiwen Lu | SparseMM: Head Sparsity Emerges from Visual Concept Responses in MLLMs | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal Large Language Models (MLLMs) are commonly derived by extending pre-trained Large Language Models (LLMs) with visual capabilities. In this work, we investigate how MLLMs process visual inputs by analyzing their attention mechanisms. We reveal a surprising sparsity phenomenon: only a small subset (approximately less than 5%) of attention heads in LLMs actively contribute to visual understanding, termed visual heads. To identify these heads efficiently, we design a training-free framework that quantifies head-level visual relevance through targeted response analysis. Building on this discovery, we introduce SparseMM, a KV-Cache optimization strategy that allocates asymmetric computation budgets to heads in LLMs based on their visual scores, leveraging the sparity of visual heads for accelerating the inference of MLLMs. Compared with prior KV-Cache acceleration methods that ignore the particularity of visual, SparseMM prioritizes stress and retaining visual semantics during decoding. Extensive evaluations across mainstream multimodal benchmarks demonstrate that SparseMM achieves superior accuracy-efficiency trade-offs. Notably, SparseMM delivers 1.38x real-time acceleration and 52% memory reduction during generation while maintaining performance parity on efficiency test. Our project is open sourced at https://github.com/CR400AF-A/SparseMM. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:55 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wang",
"Jiahui",
""
],
[
"Liu",
"Zuyan",
""
],
[
"Rao",
"Yongming",
""
],
[
"Lu",
"Jiwen",
""
]
] |
2506.05345 | Adrian {\L}a\'ncucki | Adrian {\L}a\'ncucki, Konrad Staniszewski, Piotr Nawrot, Edoardo M. Ponti | Inference-Time Hyper-Scaling with KV Cache Compression | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | Inference-time scaling trades efficiency for increased reasoning accuracy by generating longer or more parallel sequences. However, in Transformer LLMs, generation cost is bottlenecked by the size of the key-value (KV) cache, rather than the number of generated tokens. Hence, we explore inference-time hyper-scaling: by compressing the KV cache, we can generate more tokens within the same compute budget and further improve the accuracy of scaled inference. The success of this approach, however, hinges on the ability of compression methods to preserve accuracy even at high compression ratios. To make hyper-scaling practical, we introduce Dynamic Memory Sparsification (DMS), a novel method for sparsifying KV caches that only requires 1K training steps to achieve 8$\times$ compression, while maintaining better accuracy than training-free sparse attention. Instead of prematurely discarding cached tokens, DMS delays token eviction, implicitly merging representations and preserving critical information. We demonstrate the effectiveness of inference-time hyper-scaling with DMS on multiple families of LLMs, showing that it boosts accuracy for comparable inference runtime and memory load. For instance, we enhance Qwen-R1 32B by an average of 9.1 points on AIME 24, 7.6 on GPQA, and 9.6 on LiveCodeBench across compute budgets. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:55 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Łańcucki",
"Adrian",
""
],
[
"Staniszewski",
"Konrad",
""
],
[
"Nawrot",
"Piotr",
""
],
[
"Ponti",
"Edoardo M.",
""
]
] |
2506.05346 | Lei Hsiung | Lei Hsiung, Tianyu Pang, Yung-Chen Tang, Linyue Song, Tsung-Yi Ho, Pin-Yu Chen, Yaoqing Yang | Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets | Project Page: https://hsiung.cc/llm-similarity-risk/ | null | null | null | cs.CR cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in large language models (LLMs) have underscored their vulnerability to safety alignment jailbreaks, particularly when subjected to downstream fine-tuning. However, existing mitigation strategies primarily focus on reactively addressing jailbreak incidents after safety guardrails have been compromised, removing harmful gradients during fine-tuning, or continuously reinforcing safety alignment throughout fine-tuning. As such, they tend to overlook a critical upstream factor: the role of the original safety-alignment data. This paper therefore investigates the degradation of safety guardrails through the lens of representation similarity between upstream alignment datasets and downstream fine-tuning tasks. Our experiments demonstrate that high similarity between these datasets significantly weakens safety guardrails, making models more susceptible to jailbreaks. Conversely, low similarity between these two types of datasets yields substantially more robust models and thus reduces harmfulness score by up to 10.33%. By highlighting the importance of upstream dataset design in the building of durable safety guardrails and reducing real-world vulnerability to jailbreak attacks, these findings offer actionable insights for fine-tuning service providers. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:55 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Hsiung",
"Lei",
""
],
[
"Pang",
"Tianyu",
""
],
[
"Tang",
"Yung-Chen",
""
],
[
"Song",
"Linyue",
""
],
[
"Ho",
"Tsung-Yi",
""
],
[
"Chen",
"Pin-Yu",
""
],
[
"Yang",
"Yaoqing",
""
]
] |
2506.05347 | Benjamin Attal | Anagh Malik, Benjamin Attal, Andrew Xie, Matthew O'Toole, David B. Lindell | Neural Inverse Rendering from Propagating Light | Website: https://anaghmalik.com/InvProp/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We present the first system for physically based, neural inverse rendering from multi-viewpoint videos of propagating light. Our approach relies on a time-resolved extension of neural radiance caching -- a technique that accelerates inverse rendering by storing infinite-bounce radiance arriving at any point from any direction. The resulting model accurately accounts for direct and indirect light transport effects and, when applied to captured measurements from a flash lidar system, enables state-of-the-art 3D reconstruction in the presence of strong indirect light. Further, we demonstrate view synthesis of propagating light, automatic decomposition of captured measurements into direct and indirect components, as well as novel capabilities such as multi-view time-resolved relighting of captured scenes. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:55 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Malik",
"Anagh",
""
],
[
"Attal",
"Benjamin",
""
],
[
"Xie",
"Andrew",
""
],
[
"O'Toole",
"Matthew",
""
],
[
"Lindell",
"David B.",
""
]
] |
2506.05348 | Yifan Wang | Yifan Wang, Peishan Yang, Zhen Xu, Jiaming Sun, Zhanhua Zhang, Yong Chen, Hujun Bao, Sida Peng, Xiaowei Zhou | FreeTimeGS: Free Gaussians at Anytime and Anywhere for Dynamic Scene Reconstruction | CVPR 2025; Project page: https://zju3dv.github.io/freetimegs/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper addresses the challenge of reconstructing dynamic 3D scenes with complex motions. Some recent works define 3D Gaussian primitives in the canonical space and use deformation fields to map canonical primitives to observation spaces, achieving real-time dynamic view synthesis. However, these methods often struggle to handle scenes with complex motions due to the difficulty of optimizing deformation fields. To overcome this problem, we propose FreeTimeGS, a novel 4D representation that allows Gaussian primitives to appear at arbitrary time and locations. In contrast to canonical Gaussian primitives, our representation possesses the strong flexibility, thus improving the ability to model dynamic 3D scenes. In addition, we endow each Gaussian primitive with an motion function, allowing it to move to neighboring regions over time, which reduces the temporal redundancy. Experiments results on several datasets show that the rendering quality of our method outperforms recent methods by a large margin. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:57 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Wang",
"Yifan",
""
],
[
"Yang",
"Peishan",
""
],
[
"Xu",
"Zhen",
""
],
[
"Sun",
"Jiaming",
""
],
[
"Zhang",
"Zhanhua",
""
],
[
"Chen",
"Yong",
""
],
[
"Bao",
"Hujun",
""
],
[
"Peng",
"Sida",
""
],
[
"Zhou",
"Xiaowei",
""
]
] |
2506.05349 | Muhammad Maaz Mr | Hanoona Rasheed, Abdelrahman Shaker, Anqi Tang, Muhammad Maaz, Ming-Hsuan Yang, Salman Khan, Fahad Khan | VideoMathQA: Benchmarking Mathematical Reasoning via Multimodal Understanding in Videos | VideoMathQA Technical Report | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Mathematical reasoning in real-world video settings presents a fundamentally different challenge than in static images or text. It requires interpreting fine-grained visual information, accurately reading handwritten or digital text, and integrating spoken cues, often dispersed non-linearly over time. In such multimodal contexts, success hinges not just on perception, but on selectively identifying and integrating the right contextual details from a rich and noisy stream of content. To this end, we introduce VideoMathQA, a benchmark designed to evaluate whether models can perform such temporally extended cross-modal reasoning on videos. The benchmark spans 10 diverse mathematical domains, covering videos ranging from 10 seconds to over 1 hour. It requires models to interpret structured visual content, understand instructional narratives, and jointly ground concepts across visual, audio, and textual modalities. We employ graduate-level experts to ensure high quality, totaling over $920$ man-hours of annotation. To reflect real-world scenarios, questions are designed around three core reasoning challenges: direct problem solving, where answers are grounded in the presented question; conceptual transfer, which requires applying learned methods to new problems; and deep instructional comprehension, involving multi-step reasoning over extended explanations and partially worked-out solutions. Each question includes multi-step reasoning annotations, enabling fine-grained diagnosis of model capabilities. Through this benchmark, we highlight the limitations of existing approaches and establish a systematic evaluation framework for models that must reason, rather than merely perceive, across temporally extended and modality-rich mathematical problem settings. Our benchmark and evaluation code are available at: https://mbzuai-oryx.github.io/VideoMathQA | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:58 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Rasheed",
"Hanoona",
""
],
[
"Shaker",
"Abdelrahman",
""
],
[
"Tang",
"Anqi",
""
],
[
"Maaz",
"Muhammad",
""
],
[
"Yang",
"Ming-Hsuan",
""
],
[
"Khan",
"Salman",
""
],
[
"Khan",
"Fahad",
""
]
] |
2506.05350 | George Stoica | George Stoica, Vivek Ramanujan, Xiang Fan, Ali Farhadi, Ranjay Krishna, Judy Hoffman | Contrastive Flow Matching | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Unconditional flow-matching trains diffusion models to transport samples from a source distribution to a target distribution by enforcing that the flows between sample pairs are unique. However, in conditional settings (e.g., class-conditioned models), this uniqueness is no longer guaranteed--flows from different conditions may overlap, leading to more ambiguous generations. We introduce Contrastive Flow Matching, an extension to the flow matching objective that explicitly enforces uniqueness across all conditional flows, enhancing condition separation. Our approach adds a contrastive objective that maximizes dissimilarities between predicted flows from arbitrary sample pairs. We validate Contrastive Flow Matching by conducting extensive experiments across varying model architectures on both class-conditioned (ImageNet-1k) and text-to-image (CC3M) benchmarks. Notably, we find that training models with Contrastive Flow Matching (1) improves training speed by a factor of up to 9x, (2) requires up to 5x fewer de-noising steps and (3) lowers FID by up to 8.9 compared to training the same models with flow matching. We release our code at: https://github.com/gstoica27/DeltaFM.git. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 17:59:58 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Stoica",
"George",
""
],
[
"Ramanujan",
"Vivek",
""
],
[
"Fan",
"Xiang",
""
],
[
"Farhadi",
"Ali",
""
],
[
"Krishna",
"Ranjay",
""
],
[
"Hoffman",
"Judy",
""
]
] |
math/0405347 | Yuri Prokhorov | Yuri G. Prokhorov | On the degree of Fano threefolds with canonical Gorenstein singularities | 38 pages, latex2e, an error on the last page is corrected | Russian Acad. Sci. Sb. Math., 196(1): 81-122, 2005 | 10.1070/SM2005v196n01ABEH000873 | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider Fano threefolds $V$ with canonical Gorenstein singularities. A sharp bound $-K_V^3\le 72$ of the degree is proved. | [
{
"version": "v1",
"created": "Tue, 18 May 2004 09:34:04 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Apr 2008 16:44:22 GMT"
},
{
"version": "v3",
"created": "Thu, 05 Jun 2025 07:31:24 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Prokhorov",
"Yuri G.",
""
]
] |
math/0701235 | Jinzhu Han | Jinzhu Han | The Transformation of Sieve Function | 7 pages | null | null | null | math.GM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we introduced the theory of the sieve function transformation. Using the principle of sieve function transformation, we improved sieve method, and obtained the difference range of similar sieve function values. For this, we improved the prime theorem for arithmetic series. The theory of sieve function transformation can also be used to research many problems in number theory. | [
{
"version": "v1",
"created": "Mon, 08 Jan 2007 23:04:24 GMT"
},
{
"version": "v2",
"created": "Sun, 26 Oct 2008 09:05:21 GMT"
},
{
"version": "v3",
"created": "Sat, 14 Dec 2019 08:38:44 GMT"
},
{
"version": "v4",
"created": "Tue, 21 Jul 2020 13:17:58 GMT"
},
{
"version": "v5",
"created": "Wed, 04 Nov 2020 12:59:05 GMT"
},
{
"version": "v6",
"created": "Tue, 29 Dec 2020 09:27:13 GMT"
},
{
"version": "v7",
"created": "Thu, 29 Apr 2021 09:38:40 GMT"
},
{
"version": "v8",
"created": "Sun, 26 Dec 2021 13:29:06 GMT"
},
{
"version": "v9",
"created": "Mon, 17 Jan 2022 13:02:30 GMT"
},
{
"version": "v10",
"created": "Thu, 09 Jun 2022 06:23:29 GMT"
},
{
"version": "v11",
"created": "Thu, 28 Sep 2023 13:05:35 GMT"
},
{
"version": "v12",
"created": "Thu, 10 Apr 2025 02:16:27 GMT"
},
{
"version": "v13",
"created": "Thu, 17 Apr 2025 03:28:37 GMT"
},
{
"version": "v14",
"created": "Thu, 15 May 2025 03:13:21 GMT"
},
{
"version": "v15",
"created": "Thu, 05 Jun 2025 02:24:12 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Han",
"Jinzhu",
""
]
] |
quant-ph/9705030 | Masanao Ozawa | Masanao Ozawa (Nagoya University) | Quantum State Reduction and the Quantum Bayes Principle | 10 pages, LaTeX, no figures, to appear in Quantum Communication, Computing and Measurement (Kluwer, New York, 1997), 233-241 [Proc. 3rd Int'l Conf. on Quantum Communication and Measurement---QCM'96 (Hakone, 1996)] | Quantum Communication, Computing and Measurement (Kluwer, New York, 1997), 233 | 10.1007/978-1-4615-5923-8_25 | null | quant-ph | null | This paper gives new foundations of quantum state reduction without appealing to the projection postulate for the probe measurement. For this purpose, the quantum Bayes principle is formulated as the most fundamental principle for determining the state of a quantum system, and the joint probability distribution for the outcomes of local successive measurements on a noninteracting entangled system is derived without assuming the projection postulate. | [
{
"version": "v1",
"created": "Fri, 16 May 1997 09:47:09 GMT"
}
] | 2025-06-06T00:00:00 | [
[
"Ozawa",
"Masanao",
"",
"Nagoya University"
]
] |
0801.3098 | Laurence Marks | L. D. Marks and D. R. Luke | Robust Mixing for Ab-Initio Quantum Mechanical Calculations | 32 Pages, 4 figures | Physical Review B 2008 Vol. 78 Issue 7 Pages 075114-12 | 10.1103/PhysRevB.78.075114 | null | cond-mat.mtrl-sci cond-mat.str-el | null | We study the general problem of mixing for ab-initio quantum-mechanical problems. Guided by general mathematical principles and the underlying physics, we propose a multisecant form of Broydens second method for solving the self-consistent field equations of Kohn-Sham density functional theory. The algorithm is robust, requires relatively little finetuning and appears to outperform the current state of the art, converging for cases that defeat many other methods. We compare our technique to the conventional methods for problems ranging from simple to nearly pathological. | [
{
"version": "v1",
"created": "Sun, 20 Jan 2008 17:57:25 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Marks",
"L. D.",
""
],
[
"Luke",
"D. R.",
""
]
] |
1306.1138 | Attila Egri-Nagy | Attila Egri-Nagy, Chrystopher L. Nehaniv | The Attractor-Cycle Notation for Finite Transformations | 8 pages, 4 figures, this final version is implemented in SgpDec http://gap-packages.github.io/sgpdec/ | null | null | null | math.GR cs.FL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe a new notation for finite transformations. This attractor-cycle notation extends the orbit-cycle notation for permutations and builds upon existing transformation notations. How the basins of attraction of a finite transformation flow into permuted orbit cycles is visible from the notation. It gives insight into the structure of transformations and reduces the length of expressions without increasing the number of types of symbols. | [
{
"version": "v1",
"created": "Mon, 03 Jun 2013 22:31:10 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Oct 2017 13:38:50 GMT"
},
{
"version": "v3",
"created": "Wed, 12 Jun 2024 13:05:57 GMT"
},
{
"version": "v4",
"created": "Wed, 04 Jun 2025 04:28:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Egri-Nagy",
"Attila",
""
],
[
"Nehaniv",
"Chrystopher L.",
""
]
] |
1402.6774 | Jian-Xin Zhu Dr. | C.-C. Joseph Wang, Jean-Pierre Julien, A. V. Balatsky, Jian-Xin Zhu | Surface Local Impurity Scattering as a Probe for Topological Kondo Insulators | 8 pages with 4 eps figures embedded, references updated, accepted to Advanced Physics Research | null | null | LA-UR-25-20270 | cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Shortly after the discovery of topological band insulators, topological Kondo insulators (TKIs) have also been theoretically predicted. The latter has ignited revival interest in the properties of Kondo insulators. Currently, the feasibility of topological nature in SmB$_6$ has been intensively analyzed by several complementary probes. Here by starting with a minimal-orbital Anderson lattice model, we explore the local electronic structure in a Kondo insulator. We show that the two strong topological regimes sandwiching the weak topological regime give rise to a single Dirac cone, which is located near the center or corner of the surface Brillouin zone. We further find that, when a single impurity is placed on the surface, low-energy resonance states are induced in the weak scattering limit for the strong TKI regimes and the resonance level moves monotonically across the hybridization gap with the strength of impurity scattering potential; while low energy states can only be induced in the unitary scattering limit for the weak TKI regime, where the resonance level moves universally toward the center of the hybridization gap. These impurity-induced low-energy quasiparticles will lead to characteristic signatures in scanning tunneling microscopy/spectroscopy, which has recently found success in probing into exotic properties in heavy fermion systems. | [
{
"version": "v1",
"created": "Thu, 27 Feb 2014 02:57:09 GMT"
},
{
"version": "v2",
"created": "Thu, 01 Mar 2018 16:14:02 GMT"
},
{
"version": "v3",
"created": "Sat, 24 May 2025 13:21:32 GMT"
},
{
"version": "v4",
"created": "Wed, 04 Jun 2025 14:58:16 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"C. -C. Joseph",
""
],
[
"Julien",
"Jean-Pierre",
""
],
[
"Balatsky",
"A. V.",
""
],
[
"Zhu",
"Jian-Xin",
""
]
] |
1504.04852 | Mihail Cocos | Mihail Cocos | Proof of Chern conjecture for flat affine manifolds | null | null | null | null | math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove Chern's conjecture that the Euler characteristic vanishes for closed flat affine manifolds $(M,\nabla)$. The key innovation is a deformation argument for the Euler form: we construct a one-parameter family of connections $\nabla^t$ between $\nabla$ and a Levi-Civita connection, where each $\nabla^t$ is pointwise metric-compatible with respect to carefully chosen local metrics $h^{t,x} = (1-t)h^x + tg$. By lifting to $M \times [0,1]$ and exploiting the global metric-compatibility of the pullback connection $p^*\nabla^t$, we show via Stokes' theorem that the Chern-Weil integrand for $\chi(M)$ becomes exact. This approach, combined with a rigidity theorem for the associated holonomy groups (extending work of Cocos~\cite{MCocos}), forces $\chi(M) = 0$. | [
{
"version": "v1",
"created": "Sun, 19 Apr 2015 16:41:41 GMT"
},
{
"version": "v2",
"created": "Sat, 01 Aug 2015 17:06:46 GMT"
},
{
"version": "v3",
"created": "Thu, 06 Aug 2015 22:22:50 GMT"
},
{
"version": "v4",
"created": "Mon, 31 Mar 2025 13:41:30 GMT"
},
{
"version": "v5",
"created": "Fri, 11 Apr 2025 14:12:15 GMT"
},
{
"version": "v6",
"created": "Mon, 28 Apr 2025 18:16:39 GMT"
},
{
"version": "v7",
"created": "Tue, 13 May 2025 21:12:39 GMT"
},
{
"version": "v8",
"created": "Tue, 03 Jun 2025 18:58:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cocos",
"Mihail",
""
]
] |
1803.06138 | Ishfaq Ahmad Malik Mr. | Tanweer Jalal, Ishfaq Ahmad Malik | I-convergent triple difference sequence spaces defined by a sequence of modulus function | 8 pages, no figures | Journal of Nonlinear Analysis and Optimization: Theory & Applications (2019) | null | 10 (2) (2019), 49--55 | math.FA | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The main objective of this paper is to introduce classes of $I$-convergent triple difference sequence spaces, $c_{0I}^{3}(\Delta,\digamma)$, $c_{I}^{3}(\Delta,\digamma)$, $\ell_{\infty I}^{3}(\Delta,\digamma)$, $M_{I}^{3}(\Delta,\digamma)$ and $M_{0I}^{3}(\Delta,\digamma)$, by using sequence of modulii function $\digamma=(f_{pqr})$. We also study some algebraic and topological properties of these new sequence spaces. | [
{
"version": "v1",
"created": "Fri, 16 Mar 2018 10:00:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Jalal",
"Tanweer",
""
],
[
"Malik",
"Ishfaq Ahmad",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.