Peer Revenue Publish or profit

Papers

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

All of arXiv

Sunday 6 September

Computer Vision

CAM: Question Answering on Entity-Centric Videos with Continuous Extraction and Adaptive Querying

Memory facilitates question answering over long videos by extracting and retrieving facts to fit within the limited context windows of multimodal LLMs (MLLMs). Existing solutions typically extract independent memory entries from fixed-length video clips and thus cannot capture high-level semantics that need to be summarized over extended time periods, such as character traits and relations. Moreover, they rely solely on similarity-based retrieval and may fail to retrieve the fine-grained details required for question answering. To tackle these problems, we propose CAM, featuring continuous extraction for high-level semantics and adaptive querying for fine-grained details. In particular, CAM stores the entities and relations extracted from video clips in a knowledge graph. To capture the high-level semantics of each entity or relation, CAM summarizes the local subgraph of the target entity or relation once the subgraph reaches a predefined size. To retrieve the fine-grained details required for question answering, CAM supports multiple search methods, including knowledge graph traversal, video re-watching, and audio listening. It utilizes a planner-executor-verifier pipeline to adaptively compose these search methods according to question intent. Evaluations on three benchmarks show that CAM outperforms SOTA baselines and improves their accuracy by up to 23 percentage points. Code is available at https://github.com/Jake-Tian/CAM.
Read the full abstract

All authors Yizhou Tian, Zizhe Chen, Shiyuan Deng, Garry Yang, Zijie Dai, Luohao Pan, Hao Lin, Peiqi Yin, Xiao Yan, James Cheng

Differential Geometry

Short-time existence for the Schouten flow

Let $(M^n,g_0)$ be a closed smooth Riemannian manifold, with $n\geqslant 3$. We prove short–time existence and uniqueness for the "critical" Ricci–Bourguignon flow \begin{equation} \partial_t g=-2\operatorname{Ric}_g+\frac{R_g}{n-1}g\,. \end{equation} Up to a constant rescaling of time, this is the Schouten flow, since the right–hand side of the equation is $-2(n-2)$ times the Schouten tensor. At the critical parameter, however, the principal symbol of the linearization of the operator that we obtain after the usual "DeTurck modification" still has a zero eigenvalue. We resolve this degeneracy by adjoining an independent scalar unknown $r$, intended to represent the scalar curvature, and considering an "extended" system consisting of a strictly parabolic equation for the metric coupled to a transport–reaction equation for $r$ with a curvature–square forcing. The extended system belongs to the class of nonlinear composite parabolic–hyperbolic systems studied by Vol'pert and Khudyaev~\cite{VolpertKhudyaev1972}. Then, a standard compact–manifold adaptation of their theorem yields a unique smooth solution of the extended system. Finally, the difference $R_g-r$ satisfies a homogeneous linear parabolic equation, so the scalar–curvature constraint "propagates" and the original critical Ricci–Bourguignon flow is recovered. This settles the short–time existence and uniqueness problem at the critical “Schouten” value $\rho=1/(2(n-1))$, which was left open by the previous Ricci–Bourguignon theory in~\cite{CatinoEtAl2017}.
Read the full abstract

Nuclear Theory

Microscopic study of baryon stopping in low-energy heavy-ion collisions within UrQMD model

In low-energy heavy-ion collisions, baryon stopping is an important process, in which protons from the initial colliding nuclei are stopped at mid-rapidity region. Quantifying such stopping can reveal information on the properties of the nuclear medium, such as net-baryon density. Though experimental measurement of net-proton rapidity spectra provides constraints, microscopic origin of such protons is not fully accessible. Transport model studies can therefore provide deeper insight into the microscopic origin of protons transported from initial nuclei, complementing experimental measurements. This article presents an investigation of baryon stopping in minimum-bias Au+Au collisions over a wide range of beam energies, $\sqrt{s_{NN}} = 2.4-17.3$ GeV ($E_{\rm lab} = 1.23A-158A$ GeV). Final-state protons are classified based on their origin by analyzing their interaction history in the UrQMD model. A significant fraction of transported protons at mid-rapidity originates from initial neutrons rather than initial protons, referred to as isospin-converted protons, and their contribution as a function of collision energy and centrality is quantified. Anisotropic flow coefficients of different proton categories are estimated and compared with experimental measurements. Furthermore, a comparison between the $\pi^{-}/\pi^{+}$ ratio and isospin conversion rate is performed. The latter is further compared with the $\alpha$ parameters of the Kitazawa-Asakawa formalism, revealing a significant deviation from the chemical equilibrium assumption $\alpha_N = \alpha_\pi$ below $\sqrt{s_{NN}} \lesssim$ 10~GeV. Finally, the saturation of the isospin conversion rate is found to coincide with the onset of nuclear transparency, demonstrating that isospin randomization and baryon stopping are coupled phenomena in hadronic transport across the NICA/FAIR energy range.
Read the full abstract

Number Theory

Bounds on Whittaker Functions for $\mathrm{GL}(n)$

We prove a new estimate on Jacquet-Whittaker function for $\mathrm{GL}_n(\mathbb R)$, assuming that the Langlands parameters are purely imaginary and well-spaced. This gives an upper bound for the global sup-norm of the Whittaker function that matches the lower bound of Brumley-Templier in the exponent up to a linear error in $n$.
Read the full abstract

Multiagent Systems

CAPMAS: Capability-Based Delegation of Privileges in Multi-Agent Systems

Agentic systems require secure and efficient delegation of privileges across multiple collaborating agents. Existing approaches fall into two categories. Some propagate user identities directly to agents, obscuring accountability and creating persistent over-privilege risks that are amplified by the non-deterministic behaviour of AI agents. Others rely on continuous synchronization with a central Identity and Access Management (IAM) provider, introducing additional latency and communication overhead. We present CAPMAS, a novel architecture for secure end-to-end query execution in multi-agent systems. CAPMAS newly combines a contrastive learning-based semantic scoping pipeline that maps natural-language queries to bounded privilege sets before execution with expressive Macaroon-based tokens that enable offline, tamper-evident delegation with monotonic privilege reduction across agents. By decoupling authentication and delegation enforcement from agent reasoning, CAPMAS enables practical agentic execution while enforcing strict least-privilege guarantees. By eliminating synchronous delegation exchanges with the IAM, CAPMAS yields 30 times faster delegation operations, 2 times less delegation-oriented latency and up to 3 times lower bandwidth usage than the OAuth 2.0 Token Exchange (RFC 8693). Its semantic scoping pipeline achieves over 90% perfect privilege-bundle retrieval within 17 milliseconds on enterprise-scale API schemas containing over 3,100 endpoints, while reducing unnecessary privileges by 99.5% when compared to systems that propagate all the user's privileges to agents.
Read the full abstract

Machine Learning

Structural Entropy-Driven Graph Diffusion Generation for One-Shot Federated Graph Learning

One-shot federated graph learning (FGL) requires the server to estimate client contributions from highly compressed information, yet conventional volume-based weighting captures the amount of client data while overlooking how its connectivity is organized. In this paper, we propose SPIRE, a Structural Entropy-Driven Graph Diffusion Generation method that introduces topology-aware client differentiation into one-shot FGL. Specifically, we employ first-order degree-distribution structural entropy as a compact descriptor of degree-mass dispersion and use it to derive structural client weights, providing an inductive bias that accounts for differences in graph topology beyond data volume. On the generation side, a graph diffusion model on the server synthesizes pseudographs conditioned on the weighted client prototypes, capturing both semantic and structural information without requiring additional client-side training. The generated pseudographs are then assembled via disjoint union fusion to train a global graph neural network. Extensive experiments on seven real-world graph datasets demonstrate that SPIRE consistently outperforms conventional and one-shot FGL methods, with particularly strong gains under highly heterogeneous (non-IID) and graph-perturbed settings.
Read the full abstract

All authors Shutong Zheng, Lele Fu, Sheng Huang, Wei Yang Bryan Lim, Chuan Chen

Computation and Language

DFlow: Enabling Verifier Information Flow in Block Diffusion Speculative Decoding

Block diffusion speculative decoding improves LLM inference efficiency by proposing a block of future tokens in parallel and verifying them with a single forward pass through the target model. However, existing methods retain only the accepted prefix and discard the rejected suffix, preventing the computation spent on these positions from benefiting subsequent drafting rounds and forcing the drafter to repeatedly reconstruct representations for future tokens from scratch. We observe that rejection only determines whether a proposed token can be committed, while the verifier representations at rejected positions can still provide useful information for subsequent predictions. Based on this observation, we propose DFlow, a simple yet effective framework that enables verifier information to flow across drafting rounds. DFlow reuses the hidden states produced by the target verifier for the rejected suffix to guide subsequent drafting without additional target computation. To effectively learn this information flow across drafting rounds, we introduce a self-condition train strategy that feeds verifier representations from earlier predictions back into subsequent predictions. Experiments on Qwen3 models across diverse benchmarks demonstrate that DFlow consistently improves draft quality and acceptance length over DFlash.
Read the full abstract

All authors Yaojie Zhang, Linfeng Zhang, Bin Cui, Xupeng Miao

Multimedia

Vision-Guided Text Prompt Tuning for Multimodal Sentiment Analysis

Multimodal sentiment analysis requires effective modeling of both verbal semantics and non-verbal affective cues. A central challenge is to calibrate text-centered sentiment understanding with visual facial evidence in a controlled, adaptive, and parameter-efficient manner. Text usually serves as the semantic anchor, whereas visual cues provide complementary evidence for ambiguous or implicit expressions; however, indiscriminate fusion may introduce visual noise and distort textual semantics. Moreover, fully fine-tuning large visual and textual encoders is costly and prone to overfitting on limited and scenario-dependent MSA benchmarks. To address these issues, we propose Vision-Guided Text Prompt Tuning (VG-TPT), which formulates visual-text sentiment modeling as controllable visual calibration of frozen text representations. VG-TPT injects visual affective cues into a frozen BERT encoder through layer-wise adaptive prompts, rather than relying on late-stage feature fusion or full backbone tuning. A co-guided router composes prompts from a trainable prompt bank according to both the evolving text state and the visual guidance feature, enabling sample-specific and layer-specific modulation. Experiments on CMU-MOSEI and CMU-MOSI show that VG-TPT consistently improves over text-only baselines and achieves competitive or superior performance compared with several full-modality methods, while updating only 2.4M trainable parameters. The code is available at https://github.com/ma-tubu/VG-TPT.
Read the full abstract

All authors Xiaoran Kou, Jingyi Wu, Peng Sun, Yang Liu, Hong Chen

Instrumentation and Methods for Astrophysics

Mechanisms of Electrostatic Charge Formation and Retention in Lunar Regolith

As the Artemis program advances toward the lunar south pole and permanently shadowed regions (PSRs), understanding lunar charging is increasingly important for protecting astronauts, robotic systems, instruments, and infrastructure. Persistent darkness, cryogenic temperatures, low regolith conductivity, and long charge-relaxation times may allow energetic-particle-induced charge to accumulate beneath the surface. This Technical Memorandum addresses a key unresolved question: how is electrostatic charge generated, separated, retained, and accumulated within lunar regolith? Existing models predict that solar energetic particles and galactic cosmic rays may produce subsurface electric fields approaching dielectric breakdown thresholds, but the microscopic connection between incident particles and macroscopic volumetric charge sources remains unclear. Energy deposition alone does not determine retained charge. Incident particles and their secondary particles may stop, implant, backscatter, transmit, recombine, become trapped, or escape. This memorandum therefore defines the required microscopic input as the average signed retained-charge distribution per unit depth and incident particle, resolved by particle species and energy. Combined with incident flux and energy spectra, this response provides a depth-dependent volumetric charge-source rate that can be coupled with charge continuity, conduction, dielectric relaxation, and Poisson's equation. Key uncertainties include the effects of mineralogy, grain and pore geometry, temperature, and pre-existing potential on charge retention. Particle-resolved modeling and cryogenic high-vacuum irradiation experiments are needed to constrain these processes and assess subsurface electric fields, dielectric breakdown, dust transport, contamination, and charge-mitigation requirements for sustained lunar polar exploration.
Read the full abstract

All authors Minhyeok Kim, Hyun Jung Kim, Sang H. Choi, Ashley Daeun Jung

Human-Computer Interaction

Disclosure and dissolution: explainability, AI power, and situated agency in understanding

This paper interrogates the political and philosophical stakes of AI through the lens of cyborg theory cosmotechnics and glitch feminism. It advocates for a tech-positive critically situated approach to AI as a collaborator and substrate for power relations rather than an autonomous agent of harm. By rejecting the naive pause stop narratives of the human AI binary and embracing an explainable AI XAI artistic practice within embodied intersectional and community rooted engagements AI can empower diverse voices and foster ethical creativity. It starts by examining a frontier cybersecurity AI situating its framing within AI public anxiety and arguing for a new critical stance on human navigation in an AI world.
Read the full abstract

High Energy Physics (Phenomenology)

Radiative double inverse seesaw and dark matter in an alternative gauged $U(1)_{B-L}$ model

We propose a concrete theoretical framework for the double inverse seesaw mechanism within an alternative gauged $U(1)_{B-L}$ model, where the tiny neutrino masses emerge radiatively at the loop level in accordance with the 't Hooft naturalness criterion. By assigning non-universal $B-L$ charges of $(-4, -4, 5)$ to the right-handed neutrinos, we introduce vector-like fermions and two inert singlet scalars that circulate in the loop to generate the required mass terms. The spontaneous breaking of the $U(1)_{B-L}$ symmetry leaves a remnant $Z_2$ symmetry, which naturally stabilizes these new particles as dark matter (DM) candidates. We systematically investigate both fermionic and bosonic DM scenarios, assuming their interactions are predominantly mediated by the $B-L$ gauge boson ($Z'$). Our comprehensive analysis of the relic density, direct detection, and collider bounds reveals that the fermionic DM scenario is strongly favored. In contrast, the bosonic DM via the $Z'$ portal is severely constrained and largely ruled out by the latest direct detection experiments such as LZ, PandaX-4T, and XENONnT.
Read the full abstract

Astrophysics of Galaxies

A Direct DESI–SDSS Two-Fibre Test of Aperture Bias in Optical Emission-Line Diagnostics

Fixed angular apertures sample different physical regions of nearby galaxies and can therefore bias optical emission-line diagnostics. We use 20,545 high-confidence emission-line galaxies observed by both the Dark Energy Spectroscopic Instrument (DESI) and the Sloan Digital Sky Survey (SDSS) as a two-fibre experiment, in which the DESI 1.5 arcsec fibre is nested within the SDSS 3 arcsec fibre. To isolate aperture effects from line-measurement systematics, we refit every spectrum using the same eMILES stellar-continuum and emission-line model while preserving the wavelength-dependent line-spread function of each spectrum. This common refit yields 20,200 quality-controlled matched galaxies. Relative to SDSS, the smaller DESI aperture produces small but highly significant offsets: $\Delta\log({\rm N2})=+0.0067\pm0.0002$, $\Delta\log({\rm S2})=-0.0148\pm0.0003$, $\Delta\log({\rm O3})=-0.0246\pm0.0006$, $\Delta{\rm O3N2}=-0.0327\pm0.0007$, and $\Delta\log({\rm H}\alpha/{\rm H}\beta)=-0.0195\pm0.0003$, where $\Delta$ denotes DESI minus SDSS. These offsets persist in Baldwin–Phillips–Terlevich-selected star-forming galaxies and in a near-concentric subset. Their amplitudes vary across the redshift distribution and retain a measurable dependence on apparent fibre coverage after redshift is controlled, showing that redshift and relative aperture coverage jointly modulate the aperture terms. We conclude that, for high-S/N emission-line galaxies, DESI–SDSS comparisons carry measurable, diagnostic-dependent fibre-aperture terms even when continuum modelling, emission-line fitting, and spectral-resolution treatment are held fixed.
Read the full abstract

Geometric Topology

Locally Minimal Bridge Presentations of Knots

For classical knots, many characteristics are difficult to determine from arbitrary diagrams of the knot. Bridge number is of this type. It is possible to have a bridge presentation of a knot in which the number of bridges is only locally minimal, as shown by Ozawa and Takao. Here, we give a method for beginning with an arbitrary classical knot diagram having \(n\) crossings and producing a locally minimal bridge presentation of the knot with size \(O(n^2)\). This method requires only polynomial time and space. For some classes of knots, any local minimum is also the global minimum, which was shown for the trivial knot by Otal. Because the unknot is the only knot with bridge number \(1\), unknot recognition is in \(P\).
Read the full abstract

Computer Vision

Diffuse2Seg: Diffusion Models Can Segment Anything Without Supervision

Open-world entity segmentation aims to predict masks for arbitrary objects across domains and at multiple granularities, from parts to whole objects. In this setting, SAM sets a strong standard: trained on SA-1B, comprising 11M images and over 1B carefully annotated masks, it achieves remarkable zero-shot performance. Collecting such labels is expensive and time-consuming, however, which limits how far this recipe can scale. Text-to-image diffusion models offer a way around this. Their intermediate features transfer well across perception tasks, and since object structure emerges as the model denoises a noise sample into an image conditioned on a text prompt, that structure is already encoded in these representations. They can therefore be exploited for open-world entity segmentation without retraining or supervision. Building on this observation, we introduce Diffuse2Seg, which repurposes generative diffusion models for automatic mask generation by propagating a grid of point prompts through their self-attention representations in an edge-preserving manner. Diffuse2Seg produces multi-granular instance masks and outperforms prior state-of-the-art label generators by 4.3-7.1 p.p. in AR_1000 across five domains. Training an instance segmentation model on these generated masks advances detector-free open-world segmentation by 7.4 and 7.7 p.p. on "things" and "stuff+things" datasets and surpasses the detector-based UnSAM on "stuff+things" by 2.1 p.p. in AR_1000. Finally, we show that a model trained on Diffuse2Seg labels provides a strong initialization for semi-supervised learning, outperforming its fully supervised counterpart with already 5k labeled images.
Read the full abstract

All authors Christoph Hümmer, Joachim Sicking, Fabian Hüger, Hanno Gottschalk

Computer Vision

OracleZoom: On-Policy Self-Distillation Inspired Reference-Constrained Recursive Image Super Resolution

Recursive Super-Resolution (SR) extends fixed-scale SR to extreme magnification by repeatedly feeding predictions back into the same model, analogous to zooming an image repeatedly. However, ground truth availability at every scale, especially at depth, remains challenging as the required source resolution grows geometrically, leaving deeper predictions unsupervised. We present OracleZoom, an on-policy distillation-inspired, reference-constrained framework that trains on its trajectory while carrying the last ground-truth evidence beyond the supervision boundary. Direct and cross-scale supervision constrain verifiable content, while a no-reference quality objective guides unresolved fine-scale detail. A KL-constrained pretrained latent prior limits quality-driven drift, while EMA consistency stabilizes the supervision boundary. Across seven datasets, OracleZoom achieves the state-of-the-art SR quality across zooming scales, averaging 0.713 CLIPIQA, with larger gains on deeper scales, while significantly reducing hallucinations. Code, data, and models are available at https://dipta007.github.io/OracleZoom/ .
Read the full abstract

All authors Shubhashis Roy Dipta, Sourajit Saha, Shaswati Saha, Nobin Sarwar

Machine Learning

Power Mean Estimation in Stochastic Continuous Monte Carlo Tree Search

Monte Carlo Tree Search (MCTS) has demonstrated success in online planning for deterministic environments, yet significant challenges remain in adapting it to stochastic Markov Decision Processes (MDPs), particularly in continuous state-action spaces. Existing methods, such as HOOT, which combines MCTS with the Hierarchical Optimistic Optimization (HOO) bandit strategy, address continuous spaces but rely on a logarithmic exploration bonus that lacks theoretical guarantees in non-stationary, stochastic settings. Recent advancements, such as POLY-HOOT, introduced a polynomial bonus term to achieve convergence in deterministic MDPs, though a similar theory for stochastic MDPs remains undeveloped. In this paper, we propose a novel MCTS algorithm, \Algname, designed for continuous, stochastic MDPs. \Algname integrates a power mean as a value backup operator, alongside a polynomial exploration bonus to address the non-stationarity inherent in continuous action spaces. Our theoretical analysis establishes that \Algname converges at a polynomial rate of $\mathcal{O}(n^{-\zeta})$, $\zeta \in (0,1/2)$, where \( n \) is the number of visited trajectories, thereby extending the non-asymptotic convergence guarantees of POLY-HOOT to stochastic environments. Experimental results on stochastic tasks validate our theoretical findings, demonstrating the effectiveness of \Algname in continuous, stochastic domains.
Read the full abstract

Sound

Lead Vocal Separation from Vocal Ensemble Mixtures Using Phoneme Alignment

Contemporary a cappella singing often has a lead-and-accompaniment texture, where the lead vocal (Vo) part carries the main melody and the remaining vocal parts provide accompaniment. Owing to their distinct roles, separating the Vo part from the remaining vocal parts, referred to as Vo separation, enables downstream applications such as lyric recognition and minus-one accompaniment generation for vocal ensemble music. Despite these potential applications, acoustic cues for this task are limited because the target and interfering sources are all singing voices with similar acoustic characteristics and often overlap in time, making Vo separation challenging. In this paper, we propose a Vo separation model that uses phoneme alignment of the Vo part as auxiliary information. The proposed model is based on band-split RoPE Transformer (BS-RoFormer), a state-of-the-art music source separation model, and introduces frame-level phoneme labels into its intermediate representations using feature-wise linear modulation (FiLM). Experimental results show that phoneme-alignment conditioning improves Vo separation performance over an audio-only baseline and yields larger average gains than conditioning only on Vo singing/silence activity. Further analysis suggests that the advantage of phoneme-label information is larger when fewer remaining vocal parts share the same phoneme as Vo.
Read the full abstract

All authors Yuma Narahata, Tomohiko Nakamura, Yuki Saito, Hiroshi Saruwatari

Quantum Gases

Dynamical signatures of Hardcore-Boson Supersolid on the Triangular Lattice

We investigate the dynamical signatures of the supersolid phase in the triangular-lattice hardcore Bose-Hubbard model via large-scale quantum Monte Carlo simulations and linear spin-wave theory. We reveal a distinct momentum-space separation in the spectral weights. The transverse spectrum shows gapless Goldstone modes at the $\Gamma$ and $K$ points, while the longitudinal spectrum features a gapless mode at $K$ and a roton-like minimum at $M$. Directly in the imaginary-time domain, the supersolid is unambiguously characterized by the coexistence of transverse spectral lines at $\Gamma$ and $K$. Furthermore, we show that strong quantum fluctuations induce a breakdown of the sharp quasiparticle picture, leading to asymmetric line shapes and broad continua. Our work establishes the definitive dynamical signatures of the hardcore-boson supersolid, providing key insights into analogous spin supersolids in triangular quantum magnets.
Read the full abstract

All authors Xingchuan Zhu, Qingyang Guo, Junsong Sun, Yujun Fu, Xingmin Huo, Yixuan Huang, Shiping Feng, Richard T. Scalettar, Huaiming Guo

Lattice Field Theory

ggpeps: A package for Gauged Gaussian PEPS

Lattice gauge theories are a low-energy regularization of quantum field theories. They are the basis for the development of numerical algorithms that provide insight into the non-perturbative regime of quantum field theories which is difficult to achieve through other methods. We present a python package, ggpeps, implementing gauged Gaussian projected entangled pair states (GGPEPS), specialized to study lattice gauge theories. We give a an overview of the considered systems, describe the GGPEPS ansatz and provide a usage guide for the ggpeps package.
Read the full abstract

All authors Ariel Kelman, Itay Gomelski, Jonathan Elyovich, Erez Zohar, Patrick Emonts

General Relativity and Quantum Cosmology

Cycles Without End: An Ekpyrotic Braneworld Bounce with a Kinetically Coupled Entropic Mechanism

Cyclic cosmologies replace inflation's single creation event with an eternal sequence of smooth turnarounds, but three problems have historically stood in their way: singular bounces, the blue tilt generic to single-field ekpyrotic contraction, and Tolman's observation that entropy production should make successive cycles longer and larger without bound. We resolve all three within one explicit model. A Randall-Sundrum brane with a timelike extra dimension regularizes the bounce without violating the null energy condition, and a single scalar field's potential carries the universe from a thawing dark-energy plateau through a sign-changing turnaround into a BKL-safe ekpyrotic contraction and back. We show that no correction internal to the brane's Friedmann equation, and no straight two-field trajectory, can rescue the spectral index from the generic ekpyrotic result $n_s\approx3$. We then build the minimal completion that can: a kinetically coupled entropy field, following the general mechanism of Ijjas, Lehners, and Steinhardt, whose coupling steepness has a closed-form, exact-scale-invariance value for this model's ekpyrotic potential, with only a percent-level offset needed to match the measured tilt $n_s=0.9649$. This fixes the local non-Gaussianity ($f_{\rm NL}\approx-0.02$) and tensor-to-scalar ratio ($r\approx7.6\times10^{-26}$). Matching the observed amplitude fixes the conversion coupling and brane tension to unremarkable values ($\lambda\approx5.7\times10^{-3}$, $\rho_c^{1/4}=10^{13}\,$GeV). Finally, because the brane construction is spatially flat, Tolman's entropy argument does not bound the number of cycles here: we derive $N_{\rm DE}\sim70$-80 e-folds of accelerated expansion needed each cycle to dilute the entropy produced back to a negligible density before the next contraction.
Read the full abstract

Machine Learning

Second-Order Smooth Planning with Optimal-Transport Bellman Smoothing

Planning with a generative model aims to estimate the value of a state using as few simulator calls as possible. SmoothCruiser achieves problem-independent complexity $\widetilde O(\varepsilon^{-4})$ by exploiting the smoothness of the entropy-regularized Bellman backup, but its estimator is only first-order. We show that the sample-complexity exponent of SmoothCruiser-type planners is governed by the order $\beta$ of the local Taylor remainder, giving oracle complexity $\widetilde O(\varepsilon^{-(2+2/(\beta-1))})$: the first-order case $\beta=2$ recovers SmoothCruiser, while a second-order/cubic remainder $\beta=3$ yields $\widetilde O(\varepsilon^{-3})$. We reach this regime with an optimal-transport-smoothed Bellman backup over action distributions, which has a closed form, a policy gradient, and a Lipschitz Hessian, and whose quadratic correction admits an unbiased cross-product estimator. The resulting SecondOrderSmoothCruiser achieves $\widetilde O(\varepsilon^{-3})$ oracle complexity for fixed OT parameters, and we relate the OT, entropy-regularized, and unregularized objectives through explicit regularization-bias bounds.
Read the full abstract

Machine Learning

How Does Parameter Pruning Reshape DNN Representations? An Interaction-Driven Exploration

This study focuses on the scientific problem of understanding internal factors that govern the diverse performance degradation of deep neural networks (DNNs) when different parameters are pruned. In order to explain why pruning certain parameters leads to significant performance degradation but pruning other parameters does not, we examine how the pruning operation affects the interaction patterns encoded by the DNN. We find that when we progressively increase the pruning ratio, the interaction patterns encoded by DNNs exhibit a distinct three-phase dynamics, i.e., model performance is not largely affected until the pruning operation begins to remove low-order interactions, and low-order interactions exhibit strong generalizability. Moreover, we find that the high sensitivity of DNN performance to the pruning of certain modules is attributed to whether the pruning operation removes generalizable low-order interaction patterns.
Read the full abstract

All authors Fangbo Li, Junpeng Zhang, Qihan Ren, Quanshi Zhang

Probability

Fluctuating Kinetic Theory: A Poissonian Stochastic Boltzmann Equation

We introduce a nonlinear Poissonian fluctuating Boltzmann equation whose noise encodes both the fluctuations and the path large-deviation rate function of the underlying hard-sphere gas. Unlike the Gaussian-noise equations commonly studied in fluctuating hydrodynamics, the present equation raises a new difficulty: a Poisson random measure produces jumps that may destroy the nonnegativity of the solution. To address this problem, we construct a finite-dimensional coarse-grained jump process and prove its well-posedness and nonnegativity. For each fixed mesh, this process satisfies a good path large-deviation principle. We then fix a regularized velocity cutoff and study the asymptotic behavior of the discrete rate functions as the mesh is refined. On a class of biased regular paths, their limit is the corresponding cutoff Boltzmann large-deviation rate function associated with the hard-sphere gas. This establishes the consistency of the coarse-grained fluctuating Boltzmann model with the underlying particle system at the level of path large deviations.
Read the full abstract

Combinatorics

Induced Size-Ramsey Numbers of Odd Cycles via Spread Measures

We prove that, for every $k\ge2$ and every sufficiently large odd $n$, the induced size–Ramsey number of $C_n$ satisfies \[ \widehat R_{\mathrm{ind}}(C_n;k)=e^{\Theta(k)}n. \] This improves the previous upper bound of $e^{O(k\log k)}n$ and resolves a conjecture of Bradač, Draganić and Sudakov. We obtain this result by proving a stronger form of their odd-hole gadget conjecture: we construct a graph with $e^{O(k)}$ edges whose every $k$-edge-colouring contains a monochromatic induced odd cycle of length $O(\log k)$. Our proof develops a spread-based approach to finding induced cycles. The key idea is to place spread probability weights directly on families of candidate cycles and control their possible chords through a quadratic chord score. Combining this with random sparsification and hypergraph containers yields a sparse host graph that works for every $k$-edge-colouring.
Read the full abstract

Quantum Physics

Quantum Optics of harmonic generation in the strongly driven Jaynes-Cummings-type system

We adapt the Jaynes-Cummings model to study the interface of cavity quantum electrodynamics with strong field and attosecond physics. We show how multi-photon resonances in the Jaynes- Cummings system driven by a strong low-frequency classical light field lead to the generation of highly non-classical, quantum-correlated harmonics of the classical driver. Our treatment assumes no approximations, apart from the typical Jaynes-Cummings model assumption of only a few discrete quantum modes of light. The paper is dedicated to Joseph Henry Eberly, whose remarkable research has left indelible mark on both strong field physics and quantum optics.
Read the full abstract

All authors M Suman, Sili Yi, Maria Chekhova, Misha Ivanov

Analysis of PDEs

Nonradial stability of an infinite family of self-similar blow-up profiles for the three-dimensional Keller–Segel equation

The three-dimensional parabolic-elliptic Keller-Segel equation admits infinitely many radial self-similar blow-up profiles. Earlier work established radial and nonradial stability for one explicit profile. We study the non-explicit profiles constructed by inner-outer matching and allow perturbations without any symmetry assumption. For each sufficiently large profile index, we prove that the corresponding blow-up dynamics are asymptotically stable along a Lipschitz manifold of initial data. When the blow-up center and scale are allowed to vary, the codimension of this manifold equals the number of intrinsic unstable radial modes. We derive refined inner-outer asymptotics and two-scale estimates for the profiles and analyze the nonlocal, non-self-adjoint linearized operator in each spherical harmonic sector. In the radial sector, a reduced-mass transformation leads to a local Sturm-Liouville problem and gives the exact number of negative eigenvalues. In the $l=1$ sector, we identify the translation modes and exclude any additional unstable spectrum. For every $l\ge 2$, the sharp lower bound on the real part of the spectrum is $1/4$. We then combine the spectral decomposition with modulation equations, higher-order energy estimates, and a finite-dimensional Brouwer argument to obtain the nonlinear stability result.
Read the full abstract

Multimedia

MVWeaver: A Hierarchical Music Video Generation Agent with a Learned Song-to-Visual Bridge

Music videos are an important form of audiovisual expression in contemporary culture. They translate and extend the expressive content of songs through deliberate visual design. Existing automatic music video (MV) generation systems can generate visually plausible shots, yet often struggle with long-form coherence and song-grounded visual development. We present MVWeaver, a music video generation agent that integrates hierarchical planning with a learned song-to-visual bridge that translates song understanding into executable shot plans. The MVWeaver architecture comprises a comprehensive song analysis module, a visual planner that constructs hierarchical plans, and downstream image and video generation models that render the planned content. To equip a general-purpose LLM with MV-specific song-to-visual knowledge, we learn a bridge between song analysis and visual planning from real-MV-derived supervision and curate 1,861 real-world song–MV pairs with structured song-side, MV-side, and teacher-inferred song-to-visual rationale annotations. Using these annotations, we perform LoRA-based supervised fine-tuning (SFT) of a large language model to predict song-to-visual bridges that guide hierarchical visual planning. Our experiments demonstrate stronger song-grounded visual translation, richer visual development, and greater conceptual and shot-to-shot coherence, while ablations support the benefits of learned bridge conditioning.
Read the full abstract

All authors Sifei Li, Minyan Luo, Xu Li, Guodong Qi, Xincan Wang, Hanwen Wang, Chen Zhang, Pengfei Wan, Oliver Deussen, Weiming Dong

Commutative Algebra

A Solution to Iima–Yoshino Problem 2.3

Iima and Yoshino asked for an ideal $I$ in $S=k[x_1,x_2,\ldots]$, with $\operatorname{deg} x_i=i$, and a monomial order such that $S/I\cong k[x_i:i\equiv\pm1\pmod5], \operatorname{in}(I)=(x_i^2,x_ix_{i+1}:i\geq1).$ We construct such an ideal and monomial order over every field $k$ of characteristic different from $5$ containing an element $c$ with $c^2+c=1$. The ideal has an explicit infinite homogeneous reduced Gröbner basis. A five-periodic syzygy derived from a pentagon identity proves that all basis relations belong to $I$ and supplies standard representations for the non-coprime critical pairs. Triangular elimination establishes the graded quotient isomorphism. Together, the quotient and initial ideal descriptions yield the partition form of the first Rogers-Ramanujan identity. In each weighted degree, a perfect matching in the support of the normal-form matrix gives a bijection between the two partition classes. We formalize the complex specialization in Lean 4 using Mathlib and our set-based theory of infinite Gröbner bases, including the reduced basis, the graded quotient isomorphism, and the partition-matching theorem.
Read the full abstract

All authors Junyu Guo, Hao Shen, Junqi Liu, Lihong Zhi

Computer Vision

One MLLM, One Call: Efficient Zero-Shot Vision-and-Language Navigation via Spatial-Aware Waypoints

Vision-and-Language Navigation in Continuous Environments (VLN-CE) requires an embodied agent to navigate unseen environments by following natural language instructions. Current zero-shot VLN-CE methods either rely on pre-trained waypoint predictors or require multiple queries to large models per step. To address prohibitive inference latency and computational overhead, we propose O2C-Nav, an efficient zero-shot navigation framework that calls only a single large model once per decision step. Our approach introduces a training-free structured waypoint generator and a novel abstract representation that projects sparse, history-aware candidate waypoints directly onto RGB images as visual markers. The MLLM selects a waypoint or generates a fallback target bounding box at each step, while a low-level Fast Marching Method (FMM) planner converts the selected target into an executable collision-free path. This paradigm provides the model with concrete spatial perception and explicit memory while significantly reducing the visual processing load. Extensive evaluations on the R2R-CE and RxR-CE benchmarks demonstrate that O2C-Nav outperforms current state-of-the-art zero-shot methods, highlighting its great potential for real-time robotic deployment. Code is available at https://github.com/kkpsq/O2C-Nav-Code.
Read the full abstract

All authors Shiqi Pan, Qi Zheng, Hanqin Sun, Youjian Zhang, Daquan Feng, Xu Wang

Computer Vision

Thinking with Cameras: Active Visual Reasoning via Dynamic Viewpoint Control for Surveillance Video Understanding

Large vision-language models (LVLMs) have recently achieved remarkable progress in general-purpose video understanding. However, their application to real-world surveillance remains challenging due to the lack of large-scale domain-specific datasets and the limitation of passive observation from fixed viewpoints. In surveillance scenarios, critical visual evidence can be easily missed when targets are distant, small, occluded, or move beyond the current camera view. In this work, we introduce CamVLM, a new framework for Thinking with Cameras, which enables LVLMs to actively acquire visual evidence in real-world surveillance by continuously controlling camera viewpoints. We first construct CCTV-Anomaly, a large-scale surveillance video understanding dataset containing 14,133 videos across 10 anomaly categories, with detailed captions and event annotations. We further formulate viewpoint control as an active visual perception problem and build CamTrack-53K, an object-centric viewpoint trajectory dataset for learning camera actions. Moreover, we propose a reinforcement learning based viewpoint policy optimization framework, which models camera control as a sequential decision-making process and learns long-horizon observation strategies beyond supervised trajectory imitation. Extensive experiments demonstrate that CamVLM achieves state-of-the-art performance under both passive observation and dynamic viewpoint settings, validating the effectiveness of active camera-based reasoning for surveillance video understanding. Our datasets, model, and code will be available at https://github.com/xiaozhang79/CamVLM.
Read the full abstract

All authors Xiao Zhang, Wang Zeng, Sheng Jin, Wentao Liu, Chen Qian, Shichao Kan