Peer Revenue Play credits. Real papers.

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Tuesday 29 September

Astrophysics of Galaxies

LYRA: The formation and growth of nuclear star clusters in dwarf galaxies

Joaquin Sureda, Azadeh Fattahi, Sownak Bose et al. (+5 more)

Nuclear star clusters (NSCs) are dense massive star clusters ubiquitous in galactic centres across various mass regimes, present in up to $\sim30$ per cent of dwarf galaxies with $M_\star<10^{7}\, \mathrm{M}_\odot$. Although their main growth pathways are well studied, there is still ongoing debate regarding their formation. We use the LYRA cosmological hydrodynamical simulations to study the formation and growth of NSCs in a sample of six dwarf galaxies, where we identify NSCs with masses $M_\mathrm{NSC} \sim 10^4 - 10^6 \, \mathrm{M}_\odot$ in all galaxies in our sample. The NSCs in our sample emerge at early times as a compact component, with the host galaxy growing around them. This is consistent with a scenario in which NSCs are among the earliest structures to form in a galaxy. We characterise the origin of the stars that constitute the NSCs and explore the properties of the stellar populations in each of these channels. We find that most of the NSC mass originates from in-situ star formation within the NSCs themselves. While this might appear at odds with current observations, the bulk of this star formation occurs at early times, resulting in a significant population of old and metal-poor stars, in agreement with an early NSC formation. We also identify galaxy-galaxy mergers as a significant contributor to the NSC mass, with usually one merger dominating this contribution. Our results highlight the complexity in the growth history of NSCs in dwarf galaxies and highlight the importance of considering additional growth channels for NSCs in dwarfs galaxies.
Read the full abstract

All authors Joaquin Sureda, Azadeh Fattahi, Sownak Bose, Mariya Lyubenova, Jessica E. Doppel, Thales Gutcke, Katja Fahrion, Rüdiger Pakmor

Robotics

Skill-Space Shooting for Autonomous Robot Policy Improvement

Zihang Rui, Renhao Wang, Haoxu Huang et al. (+1 more)

Robots deployed in the physical world must be able to improve beyond their initial training as they encounter new situations and failures. For this improvement to scale across tasks, it must make effective use of experience without requiring human demonstration of each correction. Recent agentic systems offer a way to reduce this reliance on human effort by using foundation models to autonomously compose learned behaviors to complete tasks. Yet completing tasks this way does not itself teach a task policy to overcome its own failures; that requires turning these behaviors into learnable corrections for the policy. Our insight is that many such corrections are familiar short behaviors, or skills: they recur across tasks and describe actions that foundation models can reason about from a scene. We introduce skill-space shooting, which uses foundation model guidance to explore corrections through these reusable skills and turn successful trials into policy improvement. Real-world experiments show repeated improvement in policies acting autonomously, while skills can also be shared to reduce the teaching needed to improve on new tasks. By making reusable skills a source of corrective supervision, skill-space shooting enables scalable and generalizable policy improvement within and across tasks. Additional results and videos at https://skill-space-shooting.github.io.
Read the full abstract

All authors Zihang Rui, Renhao Wang, Haoxu Huang, Yang Gao

Computer Vision

Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering

Jaewoo Jung, Hyeonseo Yu, Honggyu An et al. (+10 more)

Reasoning about the 3D world from multi-view images remains a fundamental challenge for Multimodal Large Language Models (MLLMs). While modern MLLMs handle single-image inputs effectively, they struggle to integrate evidence across viewpoints into a coherent 3D understanding. A growing body of work attempts to close this gap by injecting 3D awareness into MLLMs, either by boosting fine-grained pixel-level cross-view correspondence or by fusing features from 3D geometry foundation models, yet a substantial gap to human reasoning persists. In this work, we revisit human spatial reasoning, which suggests that rather than relying on fine-grained geometry cues, humans roughly identify common objects across views, infer the relative geometry between viewpoints, and assemble a coarse 3D layout of the scene. Inspired by this process, we introduce Imagine3D-LLM, an MLLM that learns to assemble a similar compact 3D representation of the scene and conditions its answer on this representation. Concretely, we append a small set of learnable summary tokens after the image tokens, decode them into a compact 3D Gaussian Splatting representation supervised by a photometric reconstruction loss, and train jointly with the standard next-token prediction objective. Notably, although only the summary tokens receive direct reconstruction supervision, this objective also induces stronger cross-frame correspondence within the LLM's underlying image features, suggesting that learning to reconstruct propagates 3D-aware signals throughout the model. As a result, Imagine3D-LLM consistently outperforms prior approaches across multiple spatial reasoning and 3D understanding benchmarks, suggesting that imagining the scene can be more effective than being told its pixel-wise geometry.
Read the full abstract

All authors Jaewoo Jung, Hyeonseo Yu, Honggyu An, Jisang Han, Mungyeom Kim, Minkyeong Jeon, Heeseong Shin, Wonjun Moon, Federico Tombari, Daniel Barath, Marc Pollefeys, Seungryong Kim, Sunghwan Hong

Machine Learning

Breakdown of Local Denoising as Semantic Speciation

Guangkuo Liu, Mert Okyay, Yifan F. Zhang et al. (+3 more)

The dynamics of generative models exhibit two apparently distinct temporal windows: a speciation window, in which a sample commits to a semantic class, and a nonlocality window, in which local context windows become insufficient for generation. Motivated by evidence of their near-concurrence in a variety of frontier models, we investigate their relationship through the spatial distribution of semantic information. Under a "common cause" hypothesis, we prove that the nonlocality window must lie in the speciation window. This hypothesis postulates that semantic labels explain a fraction of the correlations between distant tokens, a condition that is natural for many real datasets. We further give conditions under which both windows shrink to a single limiting time as system size grows, defining a "phase transition", and verify this behavior analytically in Gaussian mixtures. Together, these results identify conditions under which semantic information explains the concurrence of speciation and nonlocality, connecting two complementary perspectives on the emergence of semantic structure in generative modeling.
Read the full abstract

All authors Guangkuo Liu, Mert Okyay, Yifan F. Zhang, Fangjun Hu, Rahul Nandkishore, Xun Gao

Statistical Methodology

Proximal causal inference through cross-proxy balancing

Grace V. Ringlein, Trang Q. Nguyen, Elizabeth A. Stuart et al. (+1 more)

Proximal causal inference identifies causal effects in the presence of unmeasured confounding by drawing on two sets of proxy variables. Identification typically utilizes bridge functions, defined as solutions to integral equations. However, the mechanism by which fitted bridge functions correct for confounding bias remains opaque, offering little to interpret, inspect or stress-test. We show that the defining equation of a treatment bridge function is already a balance condition, with a cross-proxy form: the weights are functions of the treatment confounding proxies and covariates, and they balance the outcome confounding proxies and covariates (i.e., reweighting the distribution in a particular treatment arm to match the distribution across treatment arms). Several existing identification paths via a treatment bridge function can then be interpreted as providing conditions under which balance on the outcome proxies implies balance on the unobserved confounders, which we call balance propagation. Leveraging this framing, we show that estimation of the treatment bridge function is a type of balancing weight estimation. Finally, we show that an outcome-weighted estimator form can also be obtained for a large class of proximal estimators, including those that utilize an outcome bridge function. Using this framing, we provide conditions under which common estimators are numerically equivalent.
Read the full abstract

All authors Grace V. Ringlein, Trang Q. Nguyen, Elizabeth A. Stuart, Harsh Parikh

Superconductivity

Imaging phase winding in topological superconductors with a fork-tip Josephson STM

Vladislav Poliakov and Archisman Panigrahi

The topological phase winding in a chiral superconductor remains challenging to access by conventional real-space probes. We propose a scheme to probe this winding with a fork-tip Josephson STM near an impurity, where two superconducting tips form an interferometer sensitive to phase differences between two positions. The phase winding remains accessible via the fork-tip critical current even if the relative phase between the fork-tip and sample fluctuates, establishing a route towards a local phase-sensitive STM probe of topological superconductivity. We numerically compute the interference pattern to demonstrate that the magnitude and phase of the order parameter can be reconstructed as a function of angle, providing a route towards resolving the longstanding open experimental problem of directly detecting phase winding in topological superconductors.
Read the full abstract

Robotics

In-context Robot Learning Made Simple: A Democratized Recipe for Manipulation Tasks

Minxing Li, Minghao Han, Weizhi Zhao et al. (+10 more)

We study robotic in-context learning (ICL), an emerging paradigm that enables robots to infer and execute tasks from visual demonstrations. Despite its growing promise, the problem itself remains under-defined: a visual demonstration simultaneously conveys action trajectories, object semantics, manipulation affordances, spatial relations, and task goals, making it unclear what information the robot is actually expected to follow. In this work, we first provide a clear problem definition of robot ICL that explicitly defines its learning target and resolves this fundamental prompt ambiguity. Building on this definition, we develop a minimalist and reproducible ICL framework (SimpleICL) with a visual prompt encoder and a low-cost data collection protocol. Without massive pre-training or specialized data infrastructure, our framework achieves strong performance in both simulation and real-world environments. Extensive experiments further reveal several key properties of robot ICL, including semantic discrimination and task-relevant disentanglement. We will fully open-source our data and training pipeline to facilitate systematic and reproducible research on robot ICL. The project page can be found at https://simpleicl.github.io/simpleicl.
Read the full abstract

All authors Minxing Li, Minghao Han, Weizhi Zhao, Hanwen Wang, Xiangshuo Liu, Shuyao Shang, Jingxiang Zhou, Mingchao Sun, Hongyu Pan, Mu Xu, Yu Liu, Lue Fan, Zhaoxiang Zhang

Robotics

Counterfactual Video Generation Enables Scalable Humanoid Loco-Manipulation

Zihan Wang, Zhen Wu, Pieter Abbeel et al. (+6 more)

Teaching humanoids loco-manipulation skills, such as carrying diverse objects, via visual imitation is a promising path toward generalist robots. However, collecting diverse, high-quality interaction videos, such as clips that clearly show a person's full body and unoccluded interactions with objects, poses a practical barrier to scaling this approach. We propose PRISM, a real-to-sim-to-real framework that overcomes this limitation by amplifying a handful of real videos into a large, diverse training set. PRISM first generates hundreds of diverse "counterfactual" human-object interaction videos via video-to-video (V2V) generation from a few exemplar real videos. Our contact-anchored real-to-sim pipeline then reconstructs both human and object motions, retargeting this imperfect video data into physically plausible trajectories. The intra-class variability across these counterfactual videos lets us train a single policy that generalizes to unseen objects within each category. We demonstrate the full pipeline by deploying this policy on a real robot without any real-world fine-tuning. Using only onboard depth observations, our humanoid picks up, carries, and drops objects, including boxes, barrels, bins, and balls, across novel instances, sizes, and initial configurations.
Read the full abstract

All authors Zihan Wang, Zhen Wu, Pieter Abbeel, Rocky Duan, Jitendra Malik, Carmelo Sferrazza, C. Karen Liu, Guanya Shi, Angjoo Kanazawa

Statistical Mechanics

Classification of topological phases of matter in stochastic systems

Dexin Li and Evelyn Tang

Topological phases of matter support edge states protected from random perturbations, and the identification of the discrete non-spatial symmetries and dimensions that support such phases was systematically developed for quantum systems and analogous platforms. While topological phases were also discovered in stochastic systems and posited as mechanisms for key biochemical processes, a classification consistent with their Markovian constraints remains lacking, impeding generalization to new scenarios. To address this, we develop new methods including a new homotopy approach, real-space invariants and a redefinition of the point gap to classify both transient and steady states. We find robust transient states in two symmetry classes, no symmetry and pseudo-Hermiticity respectively, while topological steady states are expected only in systems without symmetry. We identify the dimensions with non-trivial phases and their group structure, creating a rigorous framework for predicting robust behavior in active and living matter.
Read the full abstract

Computer Vision

Adversarial Training for Pixel Diffusion

Xin Lin, Zhifei Zhang, Yuqian Zhou et al. (+4 more)

Pixel diffusion models generate RGB images directly, avoiding the bottleneck of an autoencoder, yet their outputs still systematically underrepresent fine-scale natural-image statistics. We show that adversarial learning provides an effective post-training correction for this deficiency. Starting from a pretrained model, we retain its original diffusion or flow-matching objective and add an adversarial loss to the predicted output at non-high-noise timesteps, leaving the model architecture and sampling procedure unchanged. To our knowledge, this is the first systematic study of adversarial post-training for pixel diffusion. Across two pixel backbones, the method jointly improves distribution fidelity, coverage, prompt alignment, and perceptual quality. We further investigate why it works. Frequency-band and power-law analyses show that the original models systematically underproduce natural-image high-frequency content, while adversarial post-training restores this missing spectral power. In contrast, perceptual loss also increases high-frequency content but sacrifices distribution fidelity and prompt alignment. Nearest-neighbor, recall, and matched no-GAN SFT controls further rule out memorization, mode dropping, and additional optimization as simple explanations. Finally, we examine the boundary of this effect. Under the tested latent diffusion configurations, the same procedure does not produce comparable joint gains and adds almost no decoded high-frequency power. These results identify direct output access to the image statistics being corrected as a key factor governing when adversarial post-training succeeds.
Read the full abstract

All authors Xin Lin, Zhifei Zhang, Yuqian Zhou, Haitian Zheng, Zhe Lin, Ming-Hsuan Yang, Truong Nguyen

Computation and Language

STEPQuant: When and Where Errors Matter in Delta-Rule Recurrent State Quantization

Bingchen Yao, Haobo Xu, Haokun Lin et al. (+6 more)

Linear attention replaces growing KV caches with fixed-size recurrent states, yet these persistent states can become a substantial memory bottleneck under concurrent serving. Directly quantizing recurrent states to low precision often leads to severe accuracy degradation, as quantization errors propagate through successive state updates. We discover that the impact of these errors depends on two complementary dimensions: temporally, errors in long-lived memory can persist across many decoding steps; spatially, errors in different key rows affect model outputs differently, while state magnitudes vary substantially along both rows and columns. Motivated by these observations, we propose STEPQuant, a spatial-temporal post-training quantization framework for Delta-rule recurrent states. STEPQuant allocates precision according to error magnitude and memory lifetime, and jointly fits key-row and value-column scales based on state distributions and key-row impact on output error. Experiments on Qwen3.8-27B and Kimi-Linear-48B-A3B-Instruct across both long- and short-generation benchmarks show that STEPQuant closely matches FP32-state accuracy under a nominal 6-bit budget and outperforms uniform INT8 in its 4-bit configuration. Integrated into SGLang with optimized GPU kernels, 6-bit STEPQuant achieves over 5x recurrent-state compression and reduces total serving memory by up to 68.7%. Our code is available at https://github.com/Dreamer-Toby/STEPQuant.
Read the full abstract

All authors Bingchen Yao, Haobo Xu, Haokun Lin, Yichen Wu, Ziyu Guo, Renrui Zhang, Zhichao Lu, Zhenan Sun, Ying Wei

Quantum Physics

When Classical Correlations Certify Entanglement Recovery

Takeru Utsumi, Yota Tachibana, Yoshifumi Nakata et al. (+4 more)

Entanglement and state distinguishability have long been central topics in quantum foundations. While each has developed into a rich subject in its own right, they can be connected through measurements in complementary bases. This connection provides insights into quantum and classical correlations and underlies many information-processing tasks, most notably quantum error correction (QEC). However, it has remained largely open whether the connection between entanglement and distinguishability extends to general measurements without assuming complementarity. We answer this question in the affirmative: the key is irreducibility induced by measurements, a much more relaxed condition than complementarity. Specifically, for POVMs satisfying an irreducibility condition, we establish a quantitative relation between the infidelity of one-sided local transformation into a maximally entangled state and the failure probability in state discrimination. We then characterize when both errors can vanish simultaneously. Our results have direct applications to entanglement distillation and QEC. Their errors can be certified by estimating classical input-output correlations from experimentally accessible local measurements, without requiring complementarity. The results also extend to quantum measurement theory. We derive novel trade-off relations that bridge two historically distinct approaches: information gain–irreversibility and observable noise–disturbance. Our connection between entanglement and distinguishability via irreducibility thus offers a common framework for entanglement distillation, QEC, and measurement trade-offs.
Read the full abstract

All authors Takeru Utsumi, Yota Tachibana, Yoshifumi Nakata, Takaya Matsuura, Ryuji Takagi, Francesco Buscemi, Masato Koashi

High Energy Physics (Theory)

Exact bispectra in strongly mixed multifield inflation

Lucas Pinol

Using the effective field theory of inflationary fluctuations extended by a massive isocurvature scalar, at unit sound speed, we compute the primordial bispectra generated by every cubic interaction of dimension four or less in the unitary gauge, with the curvature–isocurvature quadratic mixing resummed to all orders in the dimensionless mixing strength $\lambda=\rho/H$. Placing the mixing in the free Hamiltonian [...], every scale-invariant tree-level bispectrum reduces to a single Schwinger-parameter integral over independent leg kernels, and what perturbation theory calls single, double and triple exchange are contact diagrams of the resummed theory. At strong mixing we unveil genuinely new, exact multifield bispectrum shapes with no correlation to the standard shapes used in data analyses. The squeezed limit carries a cosmological collider signal with a universal frequency set by the dressed mass parameter $\mu_{\rm eff}$ at every exchange order. At weak mixing the clock coefficients are obtained in closed form and reduce [...] to the known perturbative single- and double-exchange results, exactly; the triple-exchange leading coefficient is obtained in closed form at every mass, in terms of hypergeometric functions, for the first time. At strong mixing the collider amplitude scales as $e^{\pi(\lambda-2\mu_{\rm eff})/2}$ over the whole $(\lambda,\mu_{\rm eff})$ plane. This encompasses all previously studied regimes as particular cases [...]. A first confrontation of the six exact shapes with the Planck PR4 binned bispectrum shows a preference for strong mixing, where the cosmological collider oscillations invade mildly squeezed configurations [...]. The shapes of all interactions over the whole plane [...] are released with the code that produced them, so that a template at any mixing and mass can be read off in milliseconds.
Read the full abstract

Machine Learning

LeapQuant: Efficient Linear Attention with Accurate Recurrent State Quantization

Yi Pan, Haocheng Xi, Kan Zhu et al. (+10 more)

Recent LLMs increasingly adopt hybrid designs that replace standard attention with linear attention, such as Gated DeltaNet (GDN) and Kimi Delta Attention (KDA). Although they compress the context into a fixed-size recurrent state and substantially reduce the cost of long-context processing, repeatedly reading and updating that state remains a major inference bottleneck. Quantization offers a natural way to reduce this cost, but can significantly degrade model quality, due to the accumulation of rounding errors and the presence of outlier rows and columns in the state. To address these challenges, we propose LeapQuant, a training-free method that achieves near-lossless performance under 8-bit recurrent-state quantization. First, to mitigate error accumulation, we propose per-window quantization, which leaps over a window of tokens and quantizes the state only once at its end. Within a window, outputs are computed from the fixed low-bit state together with high-precision buffered updates. Second, to reduce the error introduced by each quantization, LeapQuant retains the state's largest outliers as a few high-precision Compensator Tokens, which share the update path of real tokens. We then smooth the remaining residual before quantization to further reduce the error. Comprehensive experiments across the Qwen, Kimi, and GLM model families show that LeapQuant substantially reduces memory and compute costs during inference. With accuracy comparable to the FP32 baseline, it achieves average speedups of 2.05–3.70$\times$ at the kernel level and 1.47$\times$ for end-to-end inference on NVIDIA B200, RTX PRO 6000, and RTX 5090 GPUs.
Read the full abstract

All authors Yi Pan, Haocheng Xi, Kan Zhu, Xingyang Li, Yibo Wu, Mayank Mishra, Hongtao Zhang, William X.Zheng, Baris Kasikci, Song Han, Kurt Keutzer, Rishabh Iyer, Ion Stoica

Computer Vision

Cropland PAtteRNS: Parallel Dimensional Attention Networks and Attention to Dataset Disparity for Crop Segmentation in Satellite Imagery Time Series Data

Joseph Metcalfe, Sara Sharifzadeh and Fabio Caraffini

The landscape of satellite imagery time series datasets and boundary-pushing architectures for cropland segmentation has never been richer. However, in this gold rush, important truths are being missed on both fronts, as a drive for the most novel concepts or the largest datasets pushes finer details to the side. In this paper, we present our hybrid transformer-convolutional model, Cropland Parallel Attention and Refinement Network for Segmentation (PAtteRNS), the first model to use self-attention mechanisms separately for each of the temporal, spectral, and spatial aspects of Sentinel-2 multispectral SITS data. To achieve fully-factorised attention in our proposed model, we introduce a novel parallel transformer architecture which significantly reduces the computational complexity of triple-factorised self-attention. We validate our architecture with an in-depth ablation study, and analyse the performance of our model against state-of-the-art crop segmentation models on multiple tile-size variants of the popular PASTIS and MTLCC datasets. Our findings show our model to outperform all others in the task of crop class segmentation, verified across multiple important segmentation metrics, with especially strong performance against compared models seen in the often under-reported parcel delineation quality, for which we use the Boundary IoU metric. We also find that flawed class groupings within datasets can have a significant negative impact on model performance, and report that alternate tile-size variants of crop segmentation datasets produce results incomparable to one-another, invalidating fair comparison between model performance when trained on different tile-sizes. Based on these findings, we suggest further work is required to standardise best practices when constructing SITS crop segmentation datasets, and to enable future dynamic-tile-sizing for ideal model performance.
Read the full abstract

Robotics

Rho: A Foundation for Efficiently Adaptable VLA Models

Rho Team: Simran Bagaria, Daphne Chen, Dean Fortier et al. (+10 more)

General-purpose physical AI models must combine broad visual and linguistic capabilities with precise control across robot embodiments and efficient adaptation to downstream tasks. We introduce Rho, a family of open-weights VLA models for bimanual manipulation designed for data-light task adaptation on 3 embodiments representative of dual-arm robots across research labs and the industry – YAM Box, UR AI Trainer, and FR3 Duo. We systematically ablate Rho's action-expert architecture and training recipe, and show in controlled simulation and physical-robot experiments that embodiment midtraining improves downstream adaptation. The resulting Rho variants for YAM Box, UR AI Trainer, and FR3 Duo match or outperform existing open-weights VLAs and achieve the strongest overall performance across the tasks, embodiments, and baselines evaluated in this report. We further demonstrate the Rho model family's built-in capacity for online adaptation: an internal latent policy learns from corrective feedback to select observation-conditioned noise inputs for the frozen flow-matching action expert. With as few as 15 corrected episodes, adapting this lightweight module enables Rho to handle task situations at the fringe of its offline finetuning distribution. Together, these results position Rho as both a strong general-purpose robotic manipulation model and a practical foundation for adaptation. We release the base Rho model and the embodiment-specific checkpoints to facilitate Rho's deployment in research experiments and practical industrial use cases.
Read the full abstract

All authors Rho Team: Simran Bagaria, Daphne Chen, Dean Fortier, Jianlong Fu, Michael Harrison, Tess Hellebrekers, Neel Joshi, Andrey Kolobov, Dalton Moore, Galen Mullins, Michael Murray, Eduardo Salinas, Reuben Tan

Computer Vision

Rethinking Representations for World-Action Modeling

Haoyi Jiang, Liu Liu, Xinjiang Wang et al. (+12 more)

World-action models jointly learn robot policies and predict future observations, making the representation space an interface between control and prediction. We study the design of this space through controlled comparisons, finding that neither reconstruction fidelity nor pre-trained perceptual features alone ensure effective policy learning. These findings motivate ReWAM, a representation-centric world-action model built on pre-trained DINO features. Feature Calibration and a Temporal Representation Bottleneck organize these features into compact world states suited to dynamics modeling. Action-Grounded Representation Shaping routes only action-loss gradients to the bottleneck, thereby letting the policy shape what the representation encodes while the world model learns how it evolves. Without generative video pre-training, ReWAM achieves 93.6% success on RoboTwin 2.0. On RoboDojo, it achieves an average score of 12.29 and a success rate of 8.28% using approximately 600 hours of embodied pre-training data.
Read the full abstract

All authors Haoyi Jiang, Liu Liu, Xinjiang Wang, Zhihao Sun, Zequn Chen, Sen Wang, Xinjie Wang, Xia Chen, Jingfeng Yao, Weiheng Zhao, Shanglin Yuan, Zhizhong Su, Wei Sui, Wenyu Liu, Xinggang Wang

Game Theory

Why Last-Iterate Scale-Invariant Regret Matching Converges Linearly?

Boning Li and Longbo Huang

IREG-PRM+ normalizes the cumulative regret vector by its own norm and attains optimal regret without knowledge of the payoff scale. Run unmodified on zero-sum matrix games, it converges linearly in the last iterate, and no analysis explains why. The obstacle is that the algorithm has no fixed step size to analyze: the step size is a state variable, the inverse of a regret norm that the trajectory itself moves. Every proved linear rate for regret-matching dynamics comes from restarting or modifying the update. We identify the mechanism as norm saturation: the regret norm rises to a finite limit and freezes the step size. We prove that it always does, with an explicit bound, and that saturation forces the last-iterate Nash gap to vanish on every matrix game; pointwise convergence follows whenever the equilibrium is unique. Near a unique strictly complementary equilibrium the active support freezes in one step, and the one-round Jacobian on that support has a closed form. The last-iterate then converges linearly at a closed-form rate, provided one scale-invariant quantity stays below one: the saturated step size times the largest singular value of the value-centered payoff submatrix on the support. On the $216$-instance testbed, the $184$ instances with a resolvable limit all satisfy it. The same analysis gives a ratio certificate: observable norm-increment ratios bound the unobservable Nash-gap ratio up to a constant that enters once and does not accumulate with the iteration count. Its slope-two law holds on $96.1\%$ of the instances where the slope is measurable, and the same increment monitors progress in extensive-form games, where best-response passes can be scheduled sparsely. The code is available at https://github.com/lbn187/NormCert.
Read the full abstract

Machine Learning

A Spectral Theory of Distortion in LLM Graph Reconstruction: Sharp Bounds and Empirical Characterization

Jianru Shen

Evaluations of graph reconstruction by language models typically report a single aggregate distance between the original and the reconstructed graph. We prove that for the Wasserstein distance between Laplacian spectra such a summary is bracketed by two edge counts, the net change in edge number from below and the symmetric difference from above, each scaled by $2/n$ where $n$ is the number of vertices. The bracket is sharp: its two ends coincide exactly when the reconstruction only adds edges or only deletes them, and on that class the distance is a rescaled edge count that says nothing about which edges changed. When the ends differ, the residual between the distance and the lower end is positive only if the reconstruction both invented and lost edges, which turns it into a certificate of mixed editing computable from the reported summaries alone. We characterize these regimes in 135 reconstructions produced by three open-weight models over 45 synthetic graphs. Seventy-seven outputs are one-sided and 29 mixed outputs have $X > 0$, including cases where edge count is exactly preserved while nineteen edges were simultaneously invented and lost. The three models differ in editing policy, ranging from copying the input to attempting completion at the cost of large hallucination volume, a distinction that aggregate distortion does not reveal.
Read the full abstract

Quantum Physics

Optimal Tolerant Testing of Lindbladian Dissipation

Jinge Bao, Yiyi Cai and Zhong-Xia Shang

Quantifying dissipation is essential for controlling quantum noise and characterizing open-system dynamics. In practical experimental settings, residual noise may still persist despite efforts to suppress it, making it important to determine whether the dissipative strength remains within a prescribed tolerance threshold. We study this question for an unknown, time-independent Lindblad generator with bounded strength, where the jump operators are local but with unrestricted overlap, using only memoryless forward-evolution access. The task is to distinguish dissipative strength of at most $\varepsilon_1$ from strength of at least $\varepsilon_2$, for $0 \leq \varepsilon_1 < \varepsilon_2$, measured in the canonical dissipator's normalized Frobenius norm. We give an algorithm that solves this task in total evolution time $O(\varepsilon_2/(\varepsilon_2-\varepsilon_1)^2)$ for all nontrivial thresholds, with constant success probability, and provide a matching lower bound $\Omega(\varepsilon_2/(\varepsilon_2-\varepsilon_1)^2)$ that establishes optimality even for adaptive protocols. This extends dissipation testing to the tolerant setting and eliminates the bounded-degree requirement, making the framework applicable to a broader range of experimentally relevant settings.
Read the full abstract

Information Theory

An information identity reveals the geometry of deviation events

Akshay Balsubramani

In a deviation event, the empirical measure of $n$ independent draws lands in a set of distributions. The exponent of its probability splits, at every $n$, into two nonnegative terms. The first is the rate: $n$ times the relative entropy, from the population, of the distribution of a typical draw under the event. The second is the dependence that conditioning induces among the draws, measured by their total correlation. Their relative sizes depend on the geometry of the set. This paper focuses on this exact split and on the geometry that the interplay of its two terms reveals. The dependence vanishes exactly when the event confines every draw to a single set, and the classical rate of that constraint is exact. On a set invariant under a compact group that fixes the population and leaves no invariant set of intermediate probability, the rate vanishes; uniform convergence over a hypothesis class that such a group permutes is one. On a half-space whose threshold sits a fixed number of standard errors above the mean, the fraction of the exponent that is dependence tends to a function of the event's probability alone. On a set split into pieces, the dependence is the pieces' average plus $n$ times the Jensen-Shannon divergence of their marginals, minus the entropy of their weights. The split extends to a general reference law. When the information projection is such a law, the split gives the exact value of the expectation in the exponential change of measure to the projection.
Read the full abstract

Statistics Theory

Goodness-of-fit for distributions on metric spaces

Diego Serrano, Eduardo García-Portugués and Ingrid Van Keilegom

We propose a general goodness-of-fit framework for distributions on separable metric spaces. Under suitable identifiability conditions, probability distributions are characterized by distance profiles, which motivates their use in goodness-of-fit testing, for simple and composite null hypotheses. For composite null hypotheses, parameter estimation is incorporated via a Bahadur-type expansion, and the asymptotic distribution of the empirical process for distance profiles is obtained under the null. We define test statistics based on this empirical process and derive their asymptotic null distributions. We further study the behavior of the proposed tests under fixed and local alternatives, establishing consistency results. Multiplier bootstrap procedures are developed, and their conditional asymptotic validity is established under both simple and composite null hypotheses. The methodology is illustrated with simulation studies and real-data applications for data on the sphere, hyperboloid, and simplex.
Read the full abstract

Sound

EmoRES-TTS: Residual-Enhanced Vector Steering for Emotional Speech Generation

Kuan-Po Huang, Haohe Liu, Puyuan Peng et al. (+5 more)

Emotion-conditioned text-to-speech (TTS) models may fail to express the requested emotion reliably, and improving controllability by additional training is costly in both computation and emotion-labeled speech training data. We therefore study vector steering, a training-free approach that modifies the internal representations of a frozen model. CoCoEmo, a conventional vector steering method for emotion TTS, treats each emotion vector as an indivisible direction controlled by a single global strength, limiting adherence to the requested emotion. In this work, we first discover that an emotion vector can be decomposed into a shared component that moves speech away from neutral expression and a residual component that directs generation toward the requested emotion. Building on this finding, we propose Emotion Residual-Enhanced Steering for TTS (EmoRES), a novel method that controls the two components without retraining the backbone. On IEMOCAP, EmoRES outperforms CoCoEmo across all four objective emotion metrics on the IndexTTS-2 and CosyVoice2 backbones. Rank correlation improves by 26.13 and 12.97 percentage points, corresponding to relative gains of 118.8% and 33.1%, while emotion hit rate improves by 12.95 and 6.92 points, corresponding to relative gains of 20.1% and 9.8%. Human evaluation further shows a relative improvement up to 35.0% in the rate at which listeners correctly identified the dominant requested emotion and up to a 17.3% improvement in fidelity, while listeners prefer EmoRES for naturalness in up to 63.8% of pairwise comparisons. Component ablations further demonstrate that effective control benefits from preserving the shared component while strengthening the residual of the emotion steering vectors.
Read the full abstract

All authors Kuan-Po Huang, Haohe Liu, Puyuan Peng, Haibin Wu, Zhaoheng Ni, Hung-yi Lee, Jinwon Lee, Neha Chachra

Computer Vision

DMA$^2$: Pixel-space Distribution Matching with Adversarial and Anchor Losses

Xin Lin, Zhifei Zhang, Yuqian Zhou et al. (+6 more)

Distribution matching distillation (DMD) provides a general framework for few-step diffusion generation, but its modern text-to-image instantiations have been developed primarily around latent diffusion. It therefore overlooks key properties and design opportunities of native RGB. We revisit two DMD interfaces for pixel-space teachers. On the teacher-matching side, diagnostics show low-noise RGB matching is dominated by a local-texture cue, motivating a fixed high-noise matching band. On the real-data side, native clean-RGB outputs allow guidance from an external visual representation without traversing a decoder or sharing the heavy fake-score critic. DINO-Adv removes this critic from the adversarial gradient path and supplies local parametric patch guidance. For distribution-level guidance, we introduce AF-Loss, a parameter-free auxiliary semantic distribution-field objective designed for text-to-image DMD. It operates on detached rolling real and generated supports in the shared DINOv2 space while preserving prompt-conditioned teacher supervision. AF-Loss adds no learnable parameters or inference-time computation. Together these designs form DMA$^2$. Across DPG-Bench, GenEval, VQAScore, and COCO30K, the four-step DMA$^2$ student performs better than the 25-step teacher and evaluated few-step distillers.
Read the full abstract

All authors Xin Lin, Zhifei Zhang, Yuqian Zhou, Haitian Zheng, Shaoteng Liu, Lehan Yang, Zhe Lin, Ming-Hsuan Yang, Truong Nguyen

Computer Vision

Beyond the Timeline: Augmenting Long-Video Memory with Grounded Entity Biographies

Hui Ren, Lei Fan, Henry Pao et al. (+5 more)

Answering questions about long videos often requires connecting events involving the same objects across hours or days. Chronological descriptions and text-derived entities can leave physical identity unresolved: different objects may share a description, while observations of the same object remain disconnected across events. Retrieving relevant events therefore does not necessarily recover the "biography" of the particular entity a question concerns. To address this, we introduce Grounded Entity Biographies (GEB), a long-video memory framework that groups visually grounded observations of the same physical instance across clips into retrievable biographies while preserving the context of each moment. During question answering, the biography is retrieved alongside episodic evidence, allowing the model to follow an entity through events using identity links established during memory construction. Evaluations across four benchmarks, including day-long and week-long recordings, demonstrate improvements over prior memory frameworks in both multiple-choice and open-ended question answering. On EgoLifeQA, GEB achieves 72.0% accuracy, 4.4 percentage points above the best published result. Ablations show that grounded identity association and biography reading both contribute to the gains, which additional descriptions alone do not fully recover.
Read the full abstract

All authors Hui Ren, Lei Fan, Henry Pao, Han Guo, Zeeshan Zia, Ying Chen, Alexander Schwing, Gang Hua

Computer Vision

LongLive-Plug: Once-for-All Distillation for Video Generation

Shuai Yang, Luozhou Wang, Wei Huang et al. (+9 more)

Video diffusion models are increasingly developed into specialized models for diverse downstream tasks, and this development often includes a distillation stage, for example to accelerate sampling or to improve long-video generation. This stage is typically repeated for every specialized model. We introduce LongLive-Plug, a once-for-all distillation framework that learns reusable capabilities as LoRAs on a base model for training-free, plug-and-play deployment to compatible downstream models. These capabilities include single-pass classifier-free guidance, few-step sampling, and long-context error correction for autoregressive generation. The adapters remain reusable even when downstream models add conditioning branches, expand output channels. Despite training at a fixed guidance scale, our dedicated CFG LoRA provides text guidance control through its inference weight. Combining it with a few-step LoRA simultaneously preserves few-step generation and CFG controllability on downstream tasks. We verify training-free deployment on 54 downstream models across three backbone families and eight task categories, including world modeling, robotics, editing, and multimodal generation. The approach may support additional compatible models. Each capability can thus be distilled once per backbone family and reused without per-target retraining.
Read the full abstract

All authors Shuai Yang, Luozhou Wang, Wei Huang, ZhiFei Chen, Bohan Zhang, Xiao Fu, Qianli Ma, Chen-Hsuan Lin, Weian Mao, Bryan Chu, Song Han, Yukang Chen

Computer Vision

PowerSim: Differentiable Physics Simulation and Rendering with Power Diagrams

Trong-Tung Nguyen and Anand Bhattad

We introduce PowerSim, a method to bring physically grounded, differentiable dynamics to PowerFoam's power diagram based 3D representation. PowerSim directly couples a pre-trained PowerFoam scene to the Material Point Method (MPM) by exploiting a natural alignment between the two: the geometric and appearance properties of each primitive correspond closely to the quantities MPM already tracks as an object deforms. Consequently, simulated motion can drive the scene's geometry and appearance directly, without an auxiliary representation in between. Built on this framework, we enable a range of applications on real and synthetic scenes: (1) simulating a static scene under user interaction, (2) recovering spatially varying material fields, (3) compositing primitives from independently captured scenes into a single simulation-ready scene and (4) ray-tracing reflections that update consistently as the object deforms. Our results suggest that PowerSim excels over previous frameworks for physically grounded dynamics, while unlocking unique advantages-such as secondary ray lighting effects on dynamic scenes. Results are best viewed on our project website: https://power-sim.github.io/.
Read the full abstract

Computer Vision

FracGen: Learning How Objects Stretch and Tear with Physics-Informed Video Generation

Trong-Tung Nguyen, Jiahan Zhang and Anand Bhattad

We introduce FracGen, a fracture-aware video generation model that produces plausible, controllable fracture dynamics from a single image of an intact object, conditioned on physics signals. To train FracGen, we build FracSim, a fracture-aware simulation framework that augments material point method (MPM) simulation with a continuum damage model, producing paired fracture videos and dense, pixel-aligned physical fields at no additional cost beyond standard rendering. FracGen leverages these maps in two ways: it is trained to jointly predict them alongside RGB video, encouraging the model to capture physical state rather than surface appearance; and it is supervised with physics-informed losses that encourage consistency among the predicted maps. As a result, FracGen captures distinct material-specific fracture behavior without expensive test-time simulation or per-scene tuning, while offering fine-grained control over where an object tears, how fast the crack propagates, and how much deformation precedes failure. We further introduce a benchmark for evaluating the physical plausibility of generated fracture video, and show through extensive experiments that FracGen outperforms existing video generation baselines in both physical and visual fidelity. Results are best viewed in our project website: https://fracgen.github.io/.
Read the full abstract

High Energy Physics (Theory)

CMB Injection Bounds on Moduli Fields

Naman Agarwal, Andrew R. Frey, Ratul Mahanta et al. (+3 more)

Light scalar fields with gravitational strength interactions (moduli) are ubiquitous in string, supergravity, and extra-dimensional models. If produced in the early universe with lifetimes exceeding the epoch of recombination, slowly decaying moduli inject energetic particles into the Standard Model plasma, leaving distinct imprints on the cosmic microwave background (CMB). In this paper we systematically derive CMB injection bounds on light moduli ($10^3\ {\rm eV}\leq m_{\rm{mod}}\leq 10^8\ {\rm eV}$) across the primary cosmological production channels: vacuum misalignment, heavy scalar/inflaton decay, thermal emission from the primordial plasma including production from a Hagedorn phase of hot strings. We show that these injection bounds impose stringent constraints that are complementary to both the standard cosmological moduli problem and fifth force bounds.
Read the full abstract

All authors Naman Agarwal, Andrew R. Frey, Ratul Mahanta, Anshuman Maharana, Fernando Quevedo, Gonzalo Villa

Strongly Correlated Electrons

Neural-quantum-state based downfolding of the three-band Emery model for cuprates and nickelates

Hannah Lange, Julius F. A. Tirpitz and Annabelle Bohrdt

Understanding the physics underlying high-temperature superconductivity in cuprates and, more recently, infinite-layer nickelates has remained a central challenge in condensed-matter physics. We establish neural quantum states (NQS), specifically Hidden Fermion Determinant States (HFDS), as a scalable variational approach to the three-band Emery model of the copper- and nickel-oxide layers in these materials. After benchmarking HFDS against matrix product states on width-two geometries, we study ground states of fully two-dimensional (2D) systems of up to $10\times10$ unit cells ($300$ sites). We characterize the momentum-space distribution of dopants and find a pronounced electron-hole dichotomy similar to cuprate experiments. We further downfold the three-band model to effective single-band descriptions by constructing interacting Wannier functions. We consider a wide range of parameters – from the charge-transfer regime relevant to cuprates to the Hubbard-Mott regime of nickelates, as well as systematic scans of the charge-transfer gap that has been demonstrated to impact the critical superconducting temperatures. Across all regimes, the effective model significantly deviates from the usual Fermi-Hubbard model: The typical ratio $U/t$ is enhanced, some parameters experience a significant doping dependence, and sizable terms beyond the conventional Hubbard model are present, most notably a density-assisted hopping $t_n$. Notably, in all effective models, $t_n$ has the largest contribution to particle-hole asymmetry, rather than next-nearest-neighbor hopping contributions. The effective parameters sensitively depend on the charge-transfer energy, doping, and interaction ratios. Our results establish HFDS as an efficient tool for studying the 2D Emery model and demonstrate that single-band descriptions can require interaction terms generated by the underlying multi-band models.
Read the full abstract