Peer Revenue Returns subject to peer review

Papers

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

All of arXiv

Monday 31 August

Data Structures and Algorithms

Structural Corrections to the Bethe Approximation of the Permanent

We study deterministic approximation algorithms for the permanent of a nonnegative matrix through the Bethe permanent, an approximation computable in polynomial time. The tight analysis of Anari and Rezaei gives a universal comparison between the permanent and the Bethe permanent within a factor $(\sqrt 2)^n$. The simple example of the unweighted $4$-cycle $C_4$ (or a union of disjoint $C_4$'s) shows that this bound is tight. We show that such $4$-cycle obstructions can be identified and exploited algorithmically. Given a Bethe optimizer, our algorithm identifies nearly isolated weighted $2\times2$ blocks and peels off a vertex-disjoint family of them. If the total weighted correction is large, we can improve the Bethe approximation; if it is small, we show that the Bethe permanent is within a factor of $(\sqrt2 - \varepsilon)^n$ of the truth. Combining these facts, we obtain a deterministic polynomial time $(\sqrt2-\varepsilon)^n$-approximation algorithm for the permanent of an arbitrary nonnegative $n\times n$ matrix, where $\varepsilon>0$ is some absolute constant.
Read the full abstract

Number Theory

Large zeta sums and zeros of the Riemann zeta function

For real $t$ and $x\ge 1$, set \[ S(x,t)=\sum_{n\le x} n^{\ii t}. \] We prove an unconditional inverse theorem relating large values of $S(x,t)$, with $|t|$ large, to zeros of the Riemann zeta function near height $t$. More precisely, if $T\le |t|\le 2T$, $\exp(\sqrt{\log T})\le x\le \sqrt T$, and $|S(x,t)|=x/N$ with $N\le (\log x)^{1/100}$, then for every $cN^6\le L\le (\log x)/2$ a disk centered at $1+\ii\phi$, where $|\phi-t|\ll N$, contains at least $L/360$ zeros of $\zeta(s)$. As a consequence, a quantitative restriction on zeros in a short family of arbitrarily thin fixed windows to the left of the line $\Ree s=1$ yields $S(x,t)\ll x/(\log x)^{1/100}$ in polynomial ranges of $x$. The proof adapts the zero-forcing mechanism of Granville and Soundararajan for large character sums. In the zeta setting the spectral height is shifted by $t$, and an additional residue from the pole of $\zeta(s)$ appears in the Gaussian transform; in the range considered here that residue is exponentially small.
Read the full abstract

All authors Zikang Dong, Ruihua Wang, Weijia Wang, Hao Zhang

Econometrics

When Can We Work in Embedding Space? What Text Embeddings Preserve

When do text embeddings work as inputs to empirical analysis? Their use rests on an assumption: that we can trade text for its low-dimensional embedding, and lose little in doing so. I make that assumption precise under a generative model in which documents are mixtures of latent topics. I study two uses—clustering units in embedding space and controlling for high-dimensional text. A cluster of embeddings is a set of documents with similar topic mixtures; controlling for the embedding is equivalent to controlling for the topic mixture, so validity reduces to whether that mixture captures the confounding. In an application to 363 U.S. metropolitan areas, embedding-based clusters of LLM-generated economic descriptions recover interpretable economic archetypes and separate local employment dynamics more sharply than clustering on model residuals, or on a curated set of industry and demographic covariates.
Read the full abstract

Computation and Language

Improving Information Extraction with Learned Queries

When information extraction fails, a natural instinct is to improve the model doing it: for example, by scaling it up or refining its reasoning. In this paper, we show that another part of the pipeline matters at least as much: the queries used to elicit this information. Across four clinical benchmarks and five LLMs, improving the question design alone raises performance by 18.6 F1-score points, i.e. more than using larger extraction models. To make such question design learnable, we introduce List of Questions (LoQ), which generates document-specific question sets, and FeedQ, a feedback-driven optimization method that iteratively refines questions against extraction outcomes. The resulting optimized questions can be used to train lightweight generators: with fine-tuning, 4B-parameter models match or outperform expert-derived baselines and substantially exceed the performance of much larger untuned models. We release a dataset of 12,820 optimized questions to support a broader shift in information extraction research toward treating question design as a first-class problem.
Read the full abstract

Artificial Intelligence

Measure Before You Manage: Evaluating Agent Working Memory in Coding Agents

Agent working memory is heterogeneous. Objects such as instructions, artifacts, tool outputs, and agent-generated state play different semantic roles and exhibit different size, retention, and representation profiles. Recent work has begun to explore memory-management mechanisms that account for such heterogeneity. This work focuses on semantic heterogeneity and studies how it should shape the management and evaluation of working memory in coding agents. Across 55 archived coding-agent trajectories, we find that semantically different working-memory objects exhibit distinct retention and compression behavior. This heterogeneity motivates semantically informed memory management. We study two semantically informed strategies: an object-aware compression policy and a retrieval-based policy. Their evaluation shows that calibration gains may not transfer to held-out tasks, and that equal token budgets do not imply equal delivered context or management cost. A real-system replay further exposes serving limits that nominal budgets alone do not capture. Together, these results show why semantic structure matters for agent working memory and why evaluating memory-management strategies requires more than a nominal token budget. We organize these lessons into four levels: stored state, delivered context, management work, and task or process outcome.
Read the full abstract

All authors Le Chen, Zishen Wan, Baixi Sun, Xiaolong Ma, Chih-Hsuan Yang, Feng Yan, Sheng Di, Franck Cappello, Rajeev Thakur

Combinatorics

The toughness of random graphs

For a connected and non-complete graph $G$ of order $n$, its toughness is defined as \[ \tau(G)=\min\bigl\{|S|/c(G-S):S\subseteq V(G),\ c(G-S)>1\bigr\}, \] where $c(G-S)$ denotes the number of components of $G-S$. Let $\alpha(G)$ denote the independence number of $G$. An elementary bound on toughness is $$\tau(G)\leq\frac{n-\alpha(G)}{\alpha(G)}.$$ Fix $p\in(0,1)$, and let $G(n,p)$ be the binomial random graph on vertex set $[n]$. Set $a=\alpha(G(n,p))$. In this paper, we mainly prove that \[ \tau(G(n,p))\in\left\{\frac{n-a}{a},\frac{n-a-1}{a}\right\} \] with high probability.
Read the full abstract

Fluid Dynamics

A multiscale theory of convective momentum transport in the tropical atmosphere

A multiple-scales asymptotic framework is employed to analyze the interaction between a field of small-scale convective circulations and the large-scale tropical circulation. Closed equations governing the large-scale flow are derived, providing a systematic closure for the influence of unresolved convection. The influence of unresolved convection enters through a non-local momentum diffusion operator that provides a first-principles representation of convective momentum transport and naturally couples the vertical baroclinic modes of the large-scale circulation. The classical Matsuno–Gill model is recovered as a limiting approximation in which the derived momentum transport operator reduces to the phenomenological damping introduced by Gill, thereby placing the classical theory on a systematic first-principles foundation. Two applications of the generalized theory are considered. First, steady-state circulations driven by localized equatorially symmetric heating are examined. Convectively induced vertical-mode coupling is shown to excite additional baroclinic modes beyond those directly forced, substantially modifying the vertical structure of the Walker circulation. Second, the equatorial wave spectrum is analyzed. The resulting eigenmodes comprise mixtures of the classical equatorial wave families, causing the dispersion relations to lose their distinct branch structure and exhibit systematic frequency shifts. These results demonstrate how convectively induced momentum transport systematically modifies both steady tropical circulations and equatorial wave dynamics.
Read the full abstract

Analysis of PDEs

Revisiting Picard's proof of the Picard–Weber–Weck selection theorem

Revisiting the rationale provided by Picard in the seminal paper [13], we provide an independent, self-contained proof of the Maxwell compactness property for weak Lipschitz domains in the Euclidean setting, detouring technical complications like the differential forms setting, Gaffney's inequality for smooth domains, Calderon's extension theorem or regularity theory for PDEs, or systems thereof, on smooth domains that are used in this context. As a by-product we provide an independent proof of the classical Gaffney estimate for cubes using only $L^2$-completeness of the Fourier bases.
Read the full abstract

Computer Vision

Identity-Conditioned Latent Consistency Distillation for Face Synthesis

Diffusion models have achieved strong results in high-fidelity image synthesis, but their iterative sampling process makes large-scale generation computationally expensive. This limitation is especially relevant when generating synthetic face datasets for face recognition, where a large number of subjects with many samples in different poses, expressions, ages, etc., are required. In this work, we show that identity-conditioned face synthesis can be performed at a substantially lower computational cost by a latent Consistency Model with few iterations, without compromising image quality. For training, we distill knowledge from the foundation Diffusion Model Arc2Face (teacher) by adapting its original text-to-image pipeline to an embedding-to-face setting, replacing textual prompts with ArcFace identity embeddings. Our distilled model (student) generates identity-conditioned face images with an average inference time of 0.4819 seconds per image, compared with 2.102 seconds for Arc2Face, resulting in a 4.36$\times$ speed-up. Quantitative results, based on FID scores, show that the distilled model remains competitive with Arc2Face across all evaluation protocols. On 100k generated images, it achieves near-parity on CelebA (13.921 vs. 12.928) and outperforms the teacher on WebFace42M (9.317 vs. 9.802). Further evaluations on Synth-500 and AgeDB show a moderate performance gap for the former but comparable results for the latter. These results indicate that Arc2Face can be accelerated through task-specific latent consistency distillation while preserving high image quality for large-scale synthetic face generation. Our proposal is publicly available at https://github.com/UFPR-IPASP-PR/FaceRec-IdentityConsistency.
Read the full abstract

Computer Vision

Segmentation of Bovid Dentition Under Imperfect Annotations: A Comparative Study of Convolutional and Attention Models

Semantic segmentation decomposes an image into distinct mask regions corresponding to different object categories, such as people, cars, signs or buildings. Advances in machine learning (ML) have shifted this task away from traditional rule-based heuristics such as edge detection, towards deep neural networks (DNN) that learn to classify pixels directly. However, semantic segmentation DNNs crucially depend on expertly designed mask targets to learn from, and imperfect or misaligned masks can interfere with a model's ability to learn effectively. This paper presents a comparative study of segmentation architectures, ranging from convolutional backbones to vision transformers, applied to the B.O.V.I.D. dataset, a corpus of high-resolution bovid dental photographs paired with hand-made segmentation masks not originally designed for ML-based training. We evaluate a range of preprocessing and alignment techniques to mitigate the resulting label imperfections. We find that while these preprocessing choices have limited effect on quantitative metrics such as Dice score and mIoU, their qualitative impact on predicted masks is substantial.
Read the full abstract

All authors Keith G. Mills, Evan B. Sanders, Gregory J. Matthews, Juliet K. Brophy

Materials Science

Orbital and Spin Edelstein Effects in KTaO$_3$(110) Two-Dimensional Electron Gases

The orbital Edelstein effect converts an electric field into a non-equilibrium orbital polarization, opening new opportunities for orbitronics. Although signatures of the orbital Edelstein effect have been reported, its microscopic mechanisms and quantitative validation remain underexplored. Here, by directly linking the atomic structure of KTaO$_3$(110) two-dimensional electron gases to both their calculated and measured electronic band dispersions, we predict and provide experimental evidence for an orbital Edelstein effect that largely counterbalances its spin counterpart. Scanning transmission electron microscopy and electron energy-loss spectroscopy resolve the interfacial atomic configuration, which is used as input for density-functional calculations. Angle-resolved photoemission spectroscopy then confirms the resulting band structure, which is fitted by a tight-binding model enabling computation of the spin and orbital Edelstein responses. Harmonic magnetotransport indicates that a $\sim$20 % contribution from the orbital Edelstein response is necessary to describe the magnitude and anisotropy of the effect. Our results establish KTaO$_3$(110) as a model platform for orbitronics and demonstrate a pathway to generate and harness orbital polarization in quantum oxide systems while also offering new insights into pairing mechanisms in their superconducting state.
Read the full abstract

All authors Hugo Witt, Aravind Raji, Srijani Mallik, Börge Göbel, Luis M. Vicente-Arche, Sara Varotto, Julien Bréhin, Gerbold Ménard, Raphaël Salazar, Julien Rault, François Bertran, Patrick Le Fèvre, Isabella Boventer, Ingrid Mertig, Agnès Barthélémy, Alexandre Gloter, Annika Johansson, Nicolas Bergeal, Manuel Bibes

Operator Algebras

Obstructions to the full Hao-Ng isomorphism

We show that the Hao-Ng isomorphism for full crossed products fails to hold. More precisely, for a non-degenerate $C^*$-correspondence $X$ over a $C^*$-algebra $\mathcal{A}$ and a generalized gauge action $G \curvearrowright X$ by a locally compact group $G$, we isolate three structural obstructions for the canonical commutation $\mathcal{O}_X\rtimes G \cong \mathcal{O}_{X\rtimes G}$ between full crossed product and Cuntz-Pimsner $C^*$-algebra. The first concerns hyperrigidity, the second concerns WEP of the coefficient $C^*$-algebra $\mathcal{A}$ for trivial actions, and the third concerns the coincidence of full and reduced crossed products by the action of $G$ on $\mathcal{A}$.
Read the full abstract

Mathematical Physics

On the equivalence between the Polchinski flow and the Connes-Kreimer approaches to perturbative renormalisation

We prove a correspondence between the Polchinski flow and the Connes-Kreimer approaches to perturbative renormalisation, in the sense that the first yields the same renormalisation as the latter. More precisely, we show that an ansatz based on decorated graphs (Feynman diagrams),with a combinatorial renormalisation procedure solves the Polchinski equation. This result holds for very general Euclidean quantum field theories. We are able to derive from this the form of the renormalised potential which is, to the best of our knowledge, the most general one in the literature. The main ingredients of the proof are multiple morphism properties with respect to the renormalisation, as well as a novel duality formula for forests of decorated graphs.
Read the full abstract

Image and Video Processing

OmniRAS: Standardizing Foundation Model Training and Evaluation in Robot-Assisted Surgery

Few foundation models exist for robot-assisted surgery, partly because large robotic-surgery video corpora are difficult to assemble and existing models are evaluated mostly on laparoscopic benchmarks. Further, most existing models are evaluated on a small set of public benchmarks, mostly focused on laparoscopic surgery. We present OmniRAS, a family of 1B- and 2B-parameter V-JEPA-2.1 encoders for robot-assisted surgery, and detail their training. First, we release two densely annotated robotic-cholecystectomy datasets: OmniRAS-PR and a multi-label YT-Chole tool-verb-target task, the first triplet-style annotation for robotic cholecystectomy, together with splits, probe protocols, and an inter-rater study validating the shared phase ontology. Second, we document continued pretraining at up to 256 compute nodes with global batch 6,144 over 19 sources totaling approximately 2,650 hours of surgical video, 51% robotic, and analyze compute and data composition. Third, we evaluate against raw V-JEPA-2.1 and specialized surgical models on six tasks spanning triplet, phase, and step recognition, action segmentation, and detection, under frozen-encoder and final-four-block fine-tuning regimes. Across three seeds, this yields 254 downstream runs, including 109 with partial backbone fine-tuning. The best OmniRAS models achieve the strongest adapted results across all task families, while frozen differences are smaller.
Read the full abstract

All authors Leonardo Borgioli, Neil Getty, Wenli Xiu, Jessica Cassiani, Alvaro Ducas, Carlos Agustin Orda, Hira Waris, Fangfang Xia, Rick Stevens, Pier Cristoforo Giulianotti, Milos Zefran

Statistical Methodology

Posterior Geometry and Identifiability in Multi-Response Bayesian Calibration

Calibration under model misspecification is inherently ill-posed because calibration parameters and structural discrepancy are statistically confounded without additional assumptions. Bayesian formulations address this ambiguity through prior and covariance modeling choices, including multi-response observations and cross-source alignment. However, how these choices, together with output-dependent uncertainty, shape posterior geometry and practical parameter identifiability remains poorly understood. We present a latent-variable multi-response calibration model that combines empirical Bayes estimation of nuisance hyperparameters with a Fisher-information-based local Gaussian approximation of the calibration-parameter posterior. Closed-form expressions are derived for squared-exponential and Matérn covariance functions, revealing how multi-response correlations, latent alignment, and output-specific uncertainty influence local posterior geometry. Controlled numerical experiments show that additional responses provide complementary geometric constraints on calibration parameters, while noise modeling can improve uncertainty quantification by reducing overconfidence without necessarily improving parameter localization. A lithium-ion battery study further illustrates how the model distinguishes strongly from weakly constrained parameters. Together, these results establish expected-Fisher posterior geometry as a scalable diagnostic of local parameter identifiability in multi-response calibration.
Read the full abstract

All authors Anton van Beek, Adam. M Boyce, Will J. Dawson, James B. Robinson

Machine Learning

Does On-Policy Distillation Really Distill? From Noisy Teacher to Self-Improvement

On-policy distillation (OPD) offers dense token-level supervision as an alternative to the sparse outcome-level advantages of reinforcement learning with verifiable rewards (RLVR). However, the teacher scores student-generated trajectories that are inherently off-policy for it, so the reliability of its supervision, and hence the source of the student's improvement, remains unclear. We quantitatively analyze teacher supervision during OPD training and find substantial noise whose prevalence increases with teacher scale. Surprisingly, the student policy is insensitive to such noise, converging to comparable performance regardless of whether noisy supervision is retained or removed. Does OPD distill at all? By analyzing what drives its gains, we find that learning concentrates on low log-probability tokens, and using a single fixed negative advantage matches the performance of teacher-provided ones. This suggests that OPD works largely by suppressing low log-probability tokens, which requires no teacher. These findings motivate On-Policy Self-Adaptation (OPSA), a supervision-free method using entropy-adaptive negative advantages. It assigns stronger learning signals to high-entropy positions, suppressing tail tokens, and evenly redistributing probability mass among head tokens. Compared with the base Qwen3-1.7B, OPSA improves Avg@32 by 35.41 points on AIME24, corresponding to a 263% relative gain, and more than doubles Pass@32 across all three benchmarks. It also outperforms OPD by 16.77 points in Avg@32 on AIME24. Extensive experiments and analyses across model families and tasks further demonstrate its effectiveness and generalizability.
Read the full abstract

Machine Learning

Rotational Equivariance in Machine Learning: A Comprehensive Tutorial

Rotational symmetry is one of the most important structural principles in machine learning on 3D data. In applications ranging from physics and materials science to 3D computer vision, predictions should not depend on an arbitrary choice of coordinate frame. Rotational equivariance captures this requirement mathematically by enforcing that a rotation of the input induces a corresponding transformation of the model output. This tutorial provides a comprehensive introduction to rotational equivariance, starting from the physical and geometric intuition behind coordinate independence and building up the necessary machinery from geometric deep learning, group theory, and representation theory. We introduce message passing on Euclidean graphs, group actions and representations, spherical harmonics, Wigner matrices, tensor products, and Clebsch-Gordan decomposition, and explain how these ingredients give rise to modern equivariant architectures. We then survey the principal strategies for incorporating rotational equivariance in deep learning, including group convolutions, internal tensorial representations, and canonicalization-based methods, and discuss their practical strengths and limitations. The tutorial aims to lower the barrier to the subject by connecting the underlying mathematics to practical model design, by unifying ideas that are often expressed in different formal languages, and by helping practitioners choose among competing approaches through a clear discussion of their trade-offs.
Read the full abstract

Nuclear Theory

Electromagnetic responses induced by one- and two-body nuclear currents in inclusive electron-nucleus scattering

The longitudinal and transverse nuclear response functions encode information about the electromagnetic behavior of the nuclear target probed in inclusive electron-nucleus scattering. We present a microscopic calculation based on an independent-particle model, where a non-relativistic nuclear mean-field potential governs both the initial and final hadronic states of the interaction. We add dynamically generated short-range nucleon-nucleon correlations (SRCs), as well as meson-exchange currents (MECs) derived from one-pion exchange and intermediate $\Delta$-resonance excitation. We evaluate the contributions from one- and two-nucleon knock-out reactions to nuclear responses in inclusive electron scattering from $^{12}\mathrm{C}$ in the quasielastic and dip regions. We demonstrate that for quasielastic kinematics, SRCs quench both nuclear responses, while the MECs enhance the transverse signal, substantially improving the predicted transverse-to-longitudinal ratio. In the dip region between the quasielastic and the $\Delta$-resonance peaks, we find that the observed excess of transverse strength originates mainly from explicit two-nucleon knock-out following a $\Delta$-resonance excitation.
Read the full abstract

All authors Kajetan Niewczas, Ashish Kumar Jha, Raúl González-Jiménez, Natalie Jachowicz, Alexis Nikolakopoulos, Jan Ryckebusch

Classical Analysis and ODEs

Sharp constants for weak estimates of the Riesz Potentials

In this paper we prove the sharp weak estimates for the Riesz potentials $$\|I_{s}(f)\|_{L^{\frac{n}{n-s},\infty}}\leq \gamma_{n,s} v_n^{\frac{n-s}{n}}\frac{\Gamma(s/2)\Gamma((n+2-s)/2)}{\Gamma(n/2)}\|f\|_{L^1},~~ \text{when}~~0<s<\min\{n,2\}$$ and $$ \|I_sf\|_{L^{\frac{n}{n-s},\infty}} \leq \gamma_{n,s}v_n^{\frac{n-s}{n}}\|f\|_{L^1}, ~~\text{when}~~2\leq s<n,$$ where $\gamma_{n,s}=2^{-s}\pi^{-\frac{n}{2}}\frac{\Gamma(\frac{n-s}{2})}{\Gamma(\frac{s}{2})}$ and $v_n$ is the volume of the unit ball. The isoperimetric inequality for the Riesz capacity and the Newtonian isocapacitary inequality play a crucial role in our approach.
Read the full abstract

Instrumentation and Methods for Astrophysics

Feasibility of Capillary-Driven Orbital Liquid Mirror Telescopes

Space-based liquid mirror telescopes (LMTs) have been recently proposed to overcome the scaling limits of traditional rigid mirrors. In these new space architectures, capillary forces would shape a reflective liquid into a spherical optical surface in microgravity. However, the equilibrium, stability, and dynamic behavior of such liquid interfaces remain largely uncharacterized under orbital perturbations. This study investigates the feasibility of space LMTs by combining axisymmetric capillary with thin-film models. Solar radiation pressure and self-gravitation are found to be the dominant perturbations in Sun–Earth L2 orbits, causing equilibrium deformations that exceed the $\lambda/20$ optical threshold at apertures above 0.86~m for visible ($\lambda = 380$~nm) and 1.58~m for near-infrared observatories ($\lambda = 2.5~\mu \mathrm{m}$). Micrometeoroid impacts self-heal within seconds to minutes, settling after filling requires years to decades at large apertures, and propulsive maneuvers breach optical limits within seconds to hours. These results highlight the need for interface control in large-scale systems such as NASA's 50~m Fluidic Telescope (FLUTE). Thermocapillary actuation supports apertures of 2.3/8.0~m (VIS/NIR) at $100~\mu \mathrm{K}$ and 10.6/37.7~m at $10~\mu \mathrm{K}$ resolution. Simultaneously, parasitic Marangoni flows impose sub-mK uniformity requirements for NIR and 1–$100~\mu \mathrm{K}$ for visible operation on any capillary LMT. Departing the capillary regime by imposing a surface-normal settling acceleration, ranging from 39~mm\,s$^{-2}$ (NIR,5~m) to 17~m\,s$^{-2}$ (VIS,50~m), is the most promising route to tens of meters in aperture. This approach enables a passive implementation, effectively suppressing perturbation induced interface distortions and thermal Marangoni flows. Keywords: Liquid mirror telescope, Optics, Space
Read the full abstract

Computational Finance

Agentic Quantitative Trading: A Survey of Workflows, Systems, and Evaluation

Quantitative trading is moving from isolated predictive models toward agentic workflows that combine reasoning, tool use, memory, and feedback. This survey reviews agentic quantitative trading across five stages: factor mining, signal discovery, portfolio construction, order execution, and risk management. We further examine agentic quant trading systems through architecture, coordination, and adaptation, while comparing benchmarks across strategy construction, offline trading, live market evaluation, and reliability assessment. Our review finds that current systems remain concentrated on signal discovery, while complete integration with portfolio construction, execution, and risk control is still uncommon. Multi-agent systems also rely heavily on aggregation despite increasingly diverse workflow structures. Benchmark evidence further shows that strong model or forecasting capability does not reliably translate into trading performance under live market conditions and reliability controls. We conclude with future directions for more complete trading workflows, stronger coordination, and evaluation matched to the capability being assessed.
Read the full abstract

All authors Fengrui Hua, Hengyi Yang, Xinlei Hao, Haohan Zhang, Bokai Cao, Yiyan Qi, Jia Li, Jian Guo

Information Theory

Fast Fault-Tolerant Decoders for Hypergraph Product and Lifted-Product Codes

We design low-complexity, fault-tolerant decoders for quantum low-density parity-check (QLDPC) codes with the goal of reducing decoding latency. We target two major bottlenecks of decoding under the circuit-level noise model: (i) post-processing via order-statistics decoding (OSD), and (ii) the large number of auxiliary variable nodes commonly introduced to represent CNOT-induced correlations during syndrome extraction. Our key observation is that propagating CNOT faults (hook errors) create stabilizer-induced trapping sets (TSs) that are intrinsic to hypergraph-product (HGP) and lifted-product (LP) constructions. Therefore, instead of modeling each such fault with an explicit correlation node and relying on OSD to clean up the resulting failures, we design message-passing decoders that resolve the corresponding stabilizer-induced TSs directly. We obtain these decoders by deriving QLDPC decoders from decoders for the parent classical LDPC codes and using them collectively to correct broad families of stabilizer-induced TSs. For CNOT faults that manifest primarily as syndrome errors, we show that their effect is equivalent to a data error together with syndrome-bit measurement errors. Consequently, given repeated measurements and a decoding graph that already includes nodes representing syndrome-bit errors, no distinct variable node is needed for each CNOT fault. Using a phenomenological Tanner graph with nodes representing only data errors and syndrome-bit errors, simulations on LP codes show a significant improvement in logical error rate relative to BP+OSD. Moreover, the improved error correction relies on low-complexity finite-state iterative decoders, targeted for fast and low-latency FPGA implementation.
Read the full abstract

All authors Asit Kumar Pradhan, Nithin Raveendran, David Declercq, Bane Vasić

General Relativity and Quantum Cosmology

Bekenstein–Hod Bound: A $3.3\sigma$ Confirmation from GW250114

Black holes link gravity, thermodynamics and information via limits on the relaxation rate of a perturbed system. The Bekenstein-Hod bound imposes a minimum relaxation time at fixed temperature, but its observational test demands both black hole thermodynamic characterization and decay time measurement. Here we test this bound with GW250114, the loudest gravitational-wave signal yet observed from a binary black-hole merger. We infer the remnant temperature from pre-merger data truncated at least $10\,M$ before the peak and its longest-lived decay time from post-merger data, thereby avoiding direct reuse of the same strain samples. The ringdown frequencies and damping times are allowed to vary independently rather than being fixed to the Kerr spectrum. For the primary $t_{<}=-10\,M$ analysis, the bound is verified at $3.3-3.6\sigma$ across the focal ringdown start times, representing a substantial improvement over the $91\%$ confidence level set by GW150914. The conclusion remains robust under varied pre-merger cutoffs and explicit inclusion of the short-lived first overtone in waveform modelling. This separated-data measurement converts an information-theoretic relaxation bound into a precision test of a single astrophysical black hole.
Read the full abstract

Computation and Language

Type-Balanced Contextual Learning for Incremental Named Entity Recognition

Incremental Named Entity Recognition (INER) stands as a pivotal task in information extraction, emphasizing the successive identification of new entity types within unstructured text. Faced with the continuous influx of entity types, INER grapples with two significant challenges: the widespread issue of catastrophic forgetting and the unique shift issue of the non-entity type semantics. While pseudo-labeling-based INER methods have proven effective in addressing these challenges, a previously overlooked issue arises: the biased context problem. Our analysis shows that, in new sentences, the contextual associations of tokens representing old entity types exhibit a significantly stronger bias towards new entity types compared to their contexts in old sentences. This tendency intensifies the degradation of old knowledge while promoting the overfitting of new knowledge. To solve this biased context, we propose a Type-Balanced Contextual Learning (TBCL) method, featuring a sentence-duplet learning scheme and a contextual consistency loss. This approach offers a fresh perspective for INER through context analysis. Extensive experiments across ten INER settings on three highly recognized datasets showcase the efficacy of our TBCL method, highlighting its proficiency in resolving the biased context issue inherent in pseudo-labeling based INER approaches.
Read the full abstract

All authors Duzhen Zhang, Yahan Yu, Xiuyi Chen, Chenxing Li, Dong Yu

Computation and Language

Language-Statistical Analysis of Neural Audio Codec Tokens Across Architectures, Corpora, and Noise Conditions

Neural audio codecs (NACs) convert speech into discrete token sequences, and prior work has reported that these sequences follow language-like statistical laws. This paper analyzes the token statistics of 13 NACs spanning multi-codebook residual vector quantization (RVQ), single-codebook VQ, and non-VQ designs, evaluated on three corpora under clean, white-noise, and real-world DEMAND-noise conditions. Zipf and Heaps parameters, unigram entropy, codebook occupancy, and Jensen-Shannon divergence (JSD) are estimated from matched token samples with explicit fit-validity safeguards and family-conditional $n$-gram orders. Corpus identity explains little variance in any metric, whereas acoustic condition and quantizer meta-category dominate in a metric-dependent way, and unigram entropy is the metric most strongly associated with meta-category. Clean-to-noise JSD computed at a common unigram order is associated with mel-cepstral distortion most clearly under DEMAND noise. The collapse and explosion degradation signatures previously reported for RVQ codecs concentrate in RVQ cells under white and DEMAND noise, respectively; explosion also occurs in non-VQ codecs, and single-codebook VQ codecs shift in occupancy and distribution shape without either signature. These results provide architecture-conditioned conventions for applying language-statistical analysis to NAC tokens.
Read the full abstract

All authors Joonyong Park, Shinnosuke Takamichi, David M. Chan, Shunsuke Kando, Yuki Saito, Hiroshi Saruwatari

Machine Learning

Normalized Low-Rank Adaptation

While low-rank adaptation (LoRA) is widely used for parameter-efficient model adaptation, how to regularize its training dynamics for stable and effective optimization remains underexplored. Because LoRA initializes the up-projection to zero, its early optimization dynamics are largely governed by the down-projection. Building on this observation, we introduce Normalized Low-Rank Adaptation (NoRA), a simple yet effective method that normalizes the down-projection matrices during training. We further show that the same normalization can be applied only at initialization, improving standard LoRA without requiring repeated normalization throughout training. Across pretraining, supervised finetuning, and reinforcement learning, NoRA consistently accelerates convergence, improves performance and training stability, and mitigates catastrophic forgetting. These benefits require neither additional trainable parameters nor inference-time computation, making NoRA a simple and broadly applicable enhancement to LoRA.
Read the full abstract

All authors Jiale Kang, Ziyin Yue, Zheng Zhan, Yangyi Huang, Weiyang Liu

Computation and Language

When Does Predictor-Based RL Align with Human Perception? A Study of Subjective Rewards in Codec-Based Speech Language Models

Codec-based text-to-speech (TTS) models make language-model post-training applicable to speech generation, but it remains unclear when learned perceptual predictors can serve as reinforcement learning rewards without losing alignment with human listeners. We study this question with Group Relative Policy Optimization (GRPO) using learned rewards for anime-like speaking style, naturalness, likability, and arousal. To prevent perceptual rewards from being optimized through transcript drift, we introduce a character error rate (CER) zone constraint and compare policy optimization with Best-of-$N$ reranking under the same reward gate. Across single-reward runs, each reward primarily improves its own target metric, showing that subjective predictors are not interchangeable quality surrogates. Multi-rater A/B tests further show uneven human transfer, while a reward-gap analysis separates average transfer from within-axis calibration: signed reward gaps significantly predict listener choices in the pooled analysis, whereas residual CER gaps do not, but per-axis calibration remains heterogeneous. Best-of-8 is a strong human-level baseline and is not clearly worse than GRPO perceptually, suggesting that GRPO should be viewed as amortizing reward-selected behavior into the policy rather than uniformly outperforming reranking. These results support analyzing subjective speech rewards as predictor-axis-base tuples and provide practical diagnostics for selecting rewards before multi-reward speech post-training.
Read the full abstract

Signal Processing

XAI2CSI: Interpreting CSI with eXplainable AI for Human Activity Recognition

Wi-Fi Channel State Information (CSI) has emerged as a key enabler for device-free Human Activity Recognition (HAR), enabling low-cost, unobtrusive sensing using existing communication infrastructure. However, Deep Learning (DL) models trained on CSI data often struggle to generalize across users, environments, and device setups due to the context sensitivity of wireless propagation. Despite this challenge, limited attention has been devoted to understanding model decisions and generalization failures. This paper introduces XAI2CSI, a framework that leverages eXplainable Artificial Intelligence (XAI) to analyze DL-based CSI sensing systems. XAI2CSI employs SAGE, a model-agnostic explainability method, to quantify temporal, spectral, and spatial CSI contributions to HAR decisions under nominal and cross-context evaluations on IEEE 802.11ax data. Our analysis reveals that the considered DL model exhibits limited robustness to unseen conditions due to an over-reliance on context-specific CSI patterns, causing models to misinterpret the underlying signal dynamics when deployment conditions change. The proposed methodology and findings provide a reference framework to explore alternative solutions and guide the development of robust, transparent Wi-Fi sensing systems.
Read the full abstract

All authors Idio Guarino, Alfredo Nascita, Domenico Ciuonzo, Damiano Carra, Antonio Pescapé

Computer Vision

FaceSnap: Real-Time Personalized Lightstage Facial Performance Capture

Lightstage facial capture produces production-quality digital humans, but it is resource and labor-intensive. Multi-camera setups, hours of computation, and massive data storage create bottlenecks that hinder iterative workflows. This paper introduces FaceSnap, an end-to-end framework that streamlines capture via a two-stage approach. First, a one-time multi-view optimization from a range-of-motion sequence builds a personalized model encoding both geometry and expression-dependent appearance. This model then enables high-fidelity real-time facial performance capture from a single monocular lightstage camera, with no further multi-view capture required. FaceSnap jointly estimates geometry and dynamic 4K texture at 83 fps. The 4K texture is produced by a novel personalized residual upscaler that recovers subject-specific high-frequency detail, which generic upscalers fail to capture. FaceSnap achieves geometric accuracy competitive with full per-frame multi-view optimization while outperforming feed-forward methods trained on production-quality 3D data, all from a single camera view. Finally, we introduce Multi4D, a public benchmark for evaluating 4D facial reconstruction methods in lightstage environments, enabling topology-invariant geometric comparison across methods.
Read the full abstract

All authors Rukhshanda Hussain, Noé Artru, Emeline Got, Luiz Gustavo Hafemann, Alexandre Messier, Brandon Dearlove, Rafael M. O. Cruz, Abdallah Dib, Eric Granger

Systems and Control

A Networked SIS Epidemic–Opinion Model with Higher-Order Interactions

This paper studies a susceptible–infected–susceptible (SIS) epidemic model coupled with opinion dynamics over a network of communities with higher-order interactions. Unlike standard networked SIS models, which account only for pairwise transmission, the proposed model incorporates group-level infection mechanisms and feedback between epidemic prevalence and community opinions. We establish local stability and instability conditions for a particular healthy equilibrium, derive a sufficient condition for global exponential eradication of the infection state, and identify conditions under which higher-order interactions induce bistability in the reduced dynamics on a positively invariant synchronous set. These results characterize how higher-order interactions alter the dynamics of opinion-dependent epidemic systems. Numerical simulations illustrate the predicted eradication, bistable, and endemic regimes.
Read the full abstract

All authors Saba Samadi, José I. Caiza, Sebin Gracy, Philip E. Paré