Peer Revenue Now accepting positions

Papers

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

All of arXiv

Friday 6 November

Computation and Language

Beyond Truncation: When Can LLMs Safely Omit Generated Text?

Instruction-tuned large language models (LLMs) answer at greater length than a request needs. We ask whether a training-free controller inside the decoding loop can remove that surplus while preserving content, judged against the simplest possible baseline: truncating the model's own answer to the same length. An entropy-gated logits processor removes 25.5-34.9% of tokens on MT-Bench development questions and 19.9-34.2% on a sealed test split within pre-registered BERTScore non-inferiority margins, yet we prove and confirm by audit that its output is always a prefix of the unmodified answer: the saving is truncation. A survey of 23 training-free mechanism families finds none that selects content better than truncation without oracle labels. We therefore introduce Entailment-Licensed In-Decoding Elision (ELIDE), which, at each sentence boundary, withholds a span from the reader only when a natural-language inference model judges it entailed by the text already shown, while leaving the model's generation unchanged. ELIDE beats matched-length truncation on three model families, but reading 936 of its elisions on 670 public prompts shows that 59% harm the answer, a failure that no automatic measure detects. Scoping the licence to its discourse unit reduces the harmful share to about one elision in ten on two fresh tests.
Read the full abstract

All authors Mahmoud Alyosify, Mirna Imbabi, Eman Elkhamisi, Rahatara Ferdousi

Quantum Physics

Optimal asymptotic pure-state conversion under thermal operations

Quantum thermodynamics seeks to characterize possible state transformations under thermodynamic constraints in microscopic quantum systems. In the fully quantum regime, however, quantum coherence imposes various restrictions on state conversions, and a complete characterization of state convertibility has remained open. Here we resolve this problem for asymptotic conversions between pure states. Specifically, we show that asymptotic convertibility is fully determined by the free-energy and energy-variance ratios. Our result reveals how genuinely quantum effects quantitatively constrain state conversions in quantum thermodynamics.
Read the full abstract

Dynamical Systems

Polynomial complexity, random graphs, and strict ergodicity

We construct infinite words $w$ whose complexity functions are quadratically bounded, but whose discrete derivatives are not linearly bounded, resolving a question of Cassaigne. Moreover, we construct infinite words of polynomially bounded complexity, and subshifts of quasi-linear complexity, whose discrete derivatives do not decay compared to their complexity functions, thus exhibiting a typical exponentially-growing behavior. As an application, we give the first examples of subshifts of polynomial complexity whose Rauzy graphs do not have Benjamini-Schramm limits. Finally, we realize (up to an inevitable linear error factor) every function satisfying the necessary conditions, as the complexity function of a minimal, uniquely ergodic, cube-free subshift, strengthening previous realization results in this direction, and derive consequences for the growth of simple algebras and just-infinite semigroups satisfying nontrivial laws.
Read the full abstract

All authors Be'eri Greenfeld, Paul-Henry Leemann, Carlos Gustavo Moreira, Efim Zelmanov

Audio and Speech Processing

AccentCanvas: Activity-Guided Bidirectional Accent Conversion

Accent conversion for dubbing must change pronunciation while preserving content, voice, and source timing. Normalization alone cannot produce requested non-native accents, and matching total duration does not ensure local speech/pause alignment. We present AccentCanvas, a shared discrete diffusion model for normalization and L1-to-L2 conversion, trained on cross-paired shared-prompt recordings and synthetic pairs. Frame-level speech activity and repetition-preserving units guide local timing through synthesis. Across both tasks, AccentCanvas achieves the highest accent conversion performance (Accent Arrival) and mean listener-rated accent similarity among compared systems, with over 73% lower speech-boundary error than the best baseline for each task. These gains accompany higher L1-to-L2 WER than TokAN and Vevo-Style, highlighting a tradeoff between requested-accent resemblance and ASR-based content recognition. https://smokedindia.github.io/accentcanvas/
Read the full abstract

All authors Jaehun Kim, Kang Zhang, Sung Woo Kim, Junmo Kim, Joon Son Chung

Multiagent Systems

Phantom Consensus: Measuring Social Reasoning in Multi-Agent LLM Decision-Making

Multi-agent LLM systems increasingly aggregate judgments across model instances, but reasoning about other agents can itself introduce failure. We study a setting in which LLMs shift their public decisions after observing ambiguous social signals while reporting inaccurate estimates of what other agents privately judged, even when the group's recorded private judgments do not support the apparent consensus. We call this failure phantom consensus and study it as a pluralistic-ignorance-style pattern in controlled LLM collectives. To measure this phenomenon, we introduce PHANTOM-Bench, a benchmark that separates isolated private judgments, public decisions after ambiguous social signals, elicited estimates of the group's private judgment, transparency-based recovery, and controls for neutral public text and explicit vote-following. Across nine closed- and open-weight models, ambiguous public signals consistently degrade decision quality and can turn correct private group judgments into incorrect public collective outcomes. Public accuracy drops by as much as 47.8 percentage points, and collective rationality failures reach 47.3% in the most susceptible setting. We observe substantial errors in reported group-private beliefs alongside public-stage failures, while control conditions reveal model-dependent differences in susceptibility to ambiguous and explicit peer signals. Revealing the agents' recorded private decisions repairs many failures without revealing the gold label, although recovery varies across models. Our results show that evaluating multi-agent LLM systems requires measuring not only final agreement, but whether that agreement reflects independently measured private judgments or an apparent consensus.
Read the full abstract

All authors Disha Sheshanarayana, Rajat Subhra Pal, Manjira Sinha, Tirthankar Dasgupta

Networking

Adaptive Age of Information Optimization in Rateless Coding-Based Multicast-Enabled Sensor Networks

As the demand for real-time information in Internet of Things (IoT) and Wireless Sensor Networks (WSN) scenarios grows with the evolution of bandwidth-intensive 5G applications, multicast transmission becomes increasingly vital. This paper delves into the significant role of multicast in WSN, exploring a novel perspective of using rateless codes over the traditionally employed Hybrid automatic repeat request (HARQ) for eliminating retransmissions in a wireless multicast system. We aim to optimize data freshness, quantified by the Age of Information (AoI), and analyze strategies to minimize time-average AoI in a multicast environment with diverse channel conditions. Specifically, our policy focuses on optimizing the time-average AoI based on sensor devices' feedback. We transform the problem into a Markov Decision Process (MDP) to locate optimal and low-complexity suboptimal policies. We present the first age-minimum scheme for rateless code-based wireless multicast systems. Our numerical simulations reveal that the proposed policies, developed considering unique system structural properties, consistently surpass baseline strategies. We are thus able to preempt updates at the most beneficial time, thereby addressing the issue of the bottleneck device's adverse impact on overall performance.
Read the full abstract

Artificial Intelligence

PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals

We present PhysioWave-T, a token-efficient wavelet foundation model family for EEG, ECG, and sEMG. Dynamic scale folding recombines J+1 wavelet subbands prior to tokenization, compressing sequence length from (J+1)CS to CS tokens for J levels, C channels, and S temporal patches. Route-specific front ends and learned channel-name embeddings allow one Transformer per modality to handle diverse montages and sampling rates, while unseen setups reuse the backbone via lightweight adapters. We pre-train a 30M-parameter EEG model on over 15M windows (19–128 channels), alongside ECG (30M) and sEMG (12M) encoders. Across four EEG routes, scale folding cuts Transformer latency by 4.0–7.2x and peak memory by 6.8–12.7x over scale expansion without accuracy loss. Our EEG encoder leads 8 of 9 metrics against BENDR, BIOT, LaBraM, and EEGPT, and surpasses LUNA-L on wearable seizure detection with fewer parameters. Reliability-aware latent fusion (RALF) further integrates summary tokens from independently pre-trained encoders via sample-adaptive weights. On SeizeIT2, RALF raises event sensitivity at 11 false alarms/h from 61.4% to 74.4% and window AUPRC from 0.321 to 0.399 over the best unimodal baseline, outperforming weighted-logit and concatenation fusion across all variants. PhysioWave-T also achieves state-of-the-art sEMG accuracies and top macro-F1 on PTB-XL Super and Chapman-Shaoxing. Code and models: https://github.com/ForeverBlue816/PhysioWave.
Read the full abstract

All authors Yanlong Chen, Simone Benatti, Luca Benini, Cuntai Guan, Yawei Li

Risk Management

Portfolio-aware scenario transfer of reserve movements: evidence from Lloyd's syndicates

A claims-development movement on one portfolio may not transfer directly as a stress to another size. We estimate how to transfer such scenarios between portfolio sizes, using 632 syndicate-years from 117 Lloyd's syndicates over 2015-2024. The measured movement is the change in gross claims estimates on selected older underwriting cohorts, divided by opening gross claims outstanding plus reserves taken on in the year whose development it includes; it is not a pure measure of opening-reserve deterioration. Its dispersion falls materially with reserve size, and a size-based scale predicts held-out syndicates far better than a single market scale. But the transferred upper tail understates the smallest books: in the smallest reserve-size quartile, six of 158 held-out observations exceeded the transferred 99.5th percentile, against 0.79 expected, from three syndicates. Square-root scaling has descriptively lower held-out pinball loss at 99% and 99.5%, though more misses among the largest books. For a GBP 500m portfolio the transferred empirical 99.5th percentile is 0.34 of reserves (95% credible interval, conditional on the fitted model, 0.23-0.52), against 0.67 in the raw market pool. Both are Syndicate 5820's 2016 movement. The contribution is inspectable scenario comparability, not a calibrated capital quantile or a next-year forecast.
Read the full abstract

Computation and Language

Using an LLM to Tackle Difficult Cases in NER for Nepali Tweets

Named entity recognition (NER) is challenging for social media data in a low-resource language. LLMs often lag behind pretrained language models (PLMs) in NER performance, but they show promise in handling cases that PLMs struggle with. Using a dataset of Nepali tweets, we split the test set into "easy" and "hard" cases using token probabilities from a fine-tuned pretrained language model (PLM). We then evaluate an LLM on the hard test cases using 16 different prompt configurations. We find that selecting examples based on semantic similarity and increasing the number of examples improves F1 score by almost 10 points from the baseline configuration. Although the LLM underperforms the PLM on the full and easy test sets, it surpasses the PLM on the hard test set, resulting in an improved score when the two models are combined.
Read the full abstract

Sound

Concordance-Optimal Continuous Music Emotion Estimation: Variance Decomposition, Dual-Branch Temporal Fusion, and Conformal Prediction Intervals

Continuous music emotion recognition (MER) estimates how Arousal and Valence evolve within a song, yet the field's standard metric, pooled frame-level correlation, is dominated by between-song differences and cannot distinguish a model that tracks emotional dynamics from one that only ranks songs by their average level. We decompose annotation variance into between- and within-song components and show that mean squared error (MSE) targets the former, while windowed Lin's concordance correlation coefficient (CCC = $r \cdot C_b$) targets the latter by penalising the variance compression MSE induces. A single weight $\lambda$ controls the trade-off: on 1,764 DEAM songs, raising $\lambda$ from 0 to 1 lowers pooled Arousal $r$ from 0.843 to 0.810 while raising mean within-song $r$ from 0.204 to 0.368. Building on this objective, we present MMAS (Multimodal Music Analysis System), an open-source model that predicts Arousal and Valence every 0.5 s from raw audio. It fuses frozen MERT-v1-95M embeddings with ten symbolic features from automatic transcription through a dual-branch temporal convolutional network with cross-modal attention. Because both branches are audio-derived, one model trains jointly on DEAM and PMEmo without domain conditioning. Under song-grouped 5-fold cross-validation (2,531 songs), pooled $r$ reaches 0.801/0.685 (Arousal/Valence) on DEAM and 0.742/0.746 on PMEmo. Against a ridge baseline on identical features, within-song Arousal $r$ roughly triples (0.344 vs. 0.125; $n = 353$ songs, Wilcoxon $p = 1.7 \times 10^{-21}$, $d = 0.56$). Complete transcription failure costs at most 0.04 pooled $r$. Split-conformal intervals attain 93.7% (Arousal) and 90.0% (Valence) marginal coverage at a 95% target, with the shortfall traced to within-song dependence. MMAS is released as a pip-installable package.
Read the full abstract

All authors Mohammad Faraz Akhtar, Aditya Parashar, Ayush Jain, Madhur Gupta, Ritesh Kumar, Soubhik Chakraborty

Soft Condensed Matter

Evaporation-induced colloidal diffusiophoresis in electrolyte solutions

Non-volatile solutes accumulate near a receding drying front, generating concentration gradients that can drive directional colloidal transport by diffusiophoresis. This dynamic spatial arrangement of colloidal particles could affect the microstructure and macroscopic properties of the dried film. Here, we combine experiments and theory to investigate polystyrene particles suspended in electrolyte solutions drying within a capillary tube. By varying the electrolyte composition and ambient relative humidity, we show that evaporation-induced ion gradients drive particles toward the liquid-gas interface and enhance interfacial accumulation, whereas slower drying at higher humidity weakens the gradient and particle migration. We use the heat balance integral method to predict the transient solute field for a receding interface and couple it to particle transport to capture the observed particle trajectories and accumulation trends. These results identify background electrolytes and drying rate as coupled controls on particle distribution and provide a predictive framework to optimize formulation and processing parameters for manufacturing slurry-based materials, such as battery electrodes and turbine coatings.
Read the full abstract

All authors Dang Q. Duong, Jung Hyun Kim, Hanna Cho, Sangwoo Shin

Number Theory

Joint Approximation and the Mixed-Prime Character Group

For the integer metric defined by the square denominators of two distinct primes, we prove simultaneous approximation to zero in one prime metric and to any prescribed integer in the other. A normalized convolution argument gives a uniform central spectral bound. Random relative translations of smooth weights produce an exact positive cas cade; Fourier positivity places every normalized finer convolution below its value at zero. A far-field recursion, the growth of merged square powers, and a two-term logarithmic-form estimate give summable fac tor bounds. Consequently, for every fixed pair of distinct primes p,q, the mixed continuous character group is Z[1/(pq)]/Z: no irrational pa rameter is continuous. The mixed completion is the product of the one-prime completions. For each fixed pair this also gives a compati ble rational join and a strictly finer compatible join of the associated real TVG topologies, with the same prescribed real character group. Exact witness search terminates at every strict precision and integer target. We retain the structural, density, divisibility, sequence and finite-certificate results leading to this proof. The existence argument does not give practical witness-size or runtime bounds, and it does not assume independence or equidistribution of the two digit systems. Se lected abstract components have been checked using Lean and Mathlib; the complete analytic argument has not been formalized. The corre sponding problem for three or more primes remains separate.
Read the full abstract

Data Structures and Algorithms

Global Sinkhorn Convergence and Spectral Stability of Matrix Scaling

We give a global spectral bound for the Sinkhorn–Knopp (SK) algorithm. Let $A$ be a row-stochastic nonnegative $n\times n$ matrix admitting a finite doubly stochastic scaling $B$. Define \[ g=1-s_2(B)^2>0, \qquad D=\min\left\{\|(u,v)\|_\infty: B=\operatorname{diag}(e^u)A\operatorname{diag}(e^{-v})\right\}. \] From the standard initialization, SK achieves total marginal $\ell_1$ error $\varepsilon$ in \[ O\!\left(\frac{D+\log n+\log(1/\varepsilon)}{g}\right) \] iterations. A scale-invariant potential contracts globally according to the spectrum of $B$. The linear dependence on $D$ is necessary in the fixed-gap regime: we construct strictly positive inputs with constant input and endpoint spectral gaps, $D=\Theta(n)$, and an $\Omega(n)$ iteration lower bound at error $n^{-2}$. Our upper bound gives $O(n)$ iterations on this family. We next give spectral certificates for scalability and short scaling distance. For a row-stochastic input $A$ with positive column sums $c_j$, set \[ K=\frac{\max_jc_j}{\min_jc_j},\qquad \sigma=s_2\!\left(A\operatorname{diag}(c)^{-1/2}\right). \] The input-only condition $\sqrt K\,\sigma<1$ certifies exact scalability and polynomial scaling distortion, with $D=O(\log n)$ for fixed $K$ and a fixed positive spectral margin. Given a finite scaling $B$, the weaker condition $\sqrt K\,\sigma s_2(B)<1$ gives the same structural bounds. With a constant endpoint gap, these bounds imply logarithmic SK iteration complexity. Boundary constructions show that the strict spectral inequalities cannot uniformly be replaced by nonstrict ones.
Read the full abstract

Computer Vision

Region-aware error triage for medical image segmentation with groupwise conformal risk control

Segmentation masks are increasingly reviewed rather than redrawn, which requires knowing where a mask is likely wrong. Learned error maps rank pixels by risk, but the number of pixels to warn is chosen ad hoc, and calibrating one warning threshold over all error pixels may leave errors far from the predicted contour under-flagged. We propose RACE-Triage, a post-hoc error triage framework that leaves the segmentation mask unchanged. A risk network is trained on out-of-fold predictions from backbone-independent inputs. The predicted contour partitions each image into a boundary band and a non-boundary region, and conformal risk control is applied separately in each region to bound the expected fraction of missed errors among cases with errors there. On ISIC 2018 dermoscopy and BUSI breast ultrasound (where patient identifiers are unavailable, so independence across images is assumed), pooled control met its overall target but missed on average 31.8% and 36.4% of non-boundary errors. Regional control reduced these fractions to 19.3% and 20.9% and raised the recall of size-filtered false-negative components from 0.722 to 0.818 and from 0.611 to 0.763. The mean warned area grew by 1.6 and 4.9 percentage points, and its 95th percentile reached 23.5% and 38.5%. The effect held across three seeds and a second backbone, and area-matched heuristics did not reproduce it. On the external PH$^2$ set, frozen thresholds exceeded both targets, whereas recalibration with 100 local images brought the mean over repeated splits below them. Regional calibration improves non-boundary error coverage at a measured increase in warned area.
Read the full abstract

Fluid Dynamics

Sub-filter-scale stress and heat flux in hypersonic Couette flow and a flat-plate boundary layer

A comparative study of sub-filter-scale (SFS) quantities in hypersonic Couette flow and flat-plate boundary layer turbulence under strongly cooled-wall conditions is carried out using filtered direct numerical simulation (DNS) data to characterize the differences in SFS dynamics between internal and external wall-bounded turbulence. The two flows exhibit similar near-wall behaviour in the mean velocity profiles and SFS stress components; however, marked differences emerge in the outer layer and thermal structure. The two configurations also exhibit different thermal fields, leading to different temperature-velocity relations. The SFS behaviour shows strong sensitivity to the filtering operation: the top-hat filter yields larger SFS magnitudes, whereas the sharp spectral filter preserves a distinct spectral separation near the cutoff, leading to characteristic SFS features. Decomposition of the SFS terms shows that the residual component provides the dominant contribution to the mean SFS shear stress and heat flux. Differences between the two flows are primarily associated with the interaction terms between the resolved and residual components, which contribute with opposite signs in the two configurations, consistent with their different thermal structures. The different thermal structures also lead to distinct SFS heat-flux behaviour, including a sign change in the flat-plate boundary layer. Instantaneous SFS fields further reveal that local extrema arise from the combined contributions of multiple decomposed terms. Although these terms exhibit large local fluctuations, their net contributions to the averaged SFS quantities are relatively small. The trace of the SFS stress exhibits local negative values, reflecting local energy redistribution across the filter scale rather than a loss of physical energy.
Read the full abstract

All authors Raagvendra Singh, Sabbir Ahmed, Lian Duan, Carlo Scalo, Takahiko Toki

Functional Analysis

Disjointness preserving operators on Köthe-Banach spaces

A well-known result of Abramovich says that if $E$, $F$ are rearrangement invariant Banach spaces on a finite atomless measure space such that $E$ has an absolutely continuous norm and $E\not\subseteq F$, then there are no nonzero disjointness preserving bounded linear operators $T \colon E\to F$. We obtain an analogue of this result in the setting of general Köthe-Banach spaces. Specifically, we identify a sufficient condition on Köthe-Banach spaces $E$, $F$ on finite atomless measure spaces that implies that there are no nonzero disjointness preserving bounded linear operators $T \colon E\to F$. Consequently, by a representation theorem for regular operators due to O.~Maslyuchenko, Mykhaylyuk and the second named author, this condition implies that every regular linear operator $T \colon E\to F$ is narrow. As an application, we obtain that if $1\le \max\{p_1,p_2\}< \min\{ r_1, r_2\}\le \infty$, then every regular linear operator $T \colon L_{p_1}(L_{p_2}) \to L_{r_1}(L_{r_2})$ is narrow. The latter may be seen as an analogue of the classical Pitt compactness theorem.
Read the full abstract

Machine Learning

Accuracy Is Not Value: Choosing Weak Supervisors by What the Student Inherits

Weak supervisors are usually judged by their own errors, but what matters is what a student learns from them. For squared loss and any linear learner shared by all supervisors, we show that a supervisor's value is determined by its errors after they pass through the student's learning operator. Consequently, the inner products of the supervisors' raw errors can be compatible with every feasible value of the student's risk, and weighting supervisors by them can be arbitrarily worse than the best weighting in relative excess risk. This student-relative value is not identifiable from unlabeled data, but what is missing is small: one finite-dimensional projection of the target decides whether training on a weak aggregate helps, another decides how to weight the supervisors. Known teacher training rules, or a label-model condition stated in student-relative form, shrink it further. Clean labels close the remaining gap at a cost set by the geometry of the unresolved directions rather than by the dimension of the student's representation, with matching worst-case lower bounds. In Gaussian designs we also quantify when exact teacher structure beats ordinary stacking and when weak supervisors, combined with clean labels, beat every procedure that uses the clean labels alone. The guarantees extend to fitted ridge students and finite unlabeled data.
Read the full abstract

General Relativity and Quantum Cosmology

Bondi-Sachs formalism on past light cones and an analogue of the integral mass formula

Following a proposal that the thermodynamic aspects of general relativity can be found in the evolution of the past light cones of arbitrary worldlines, we generalize the Bondi-Sachs formalism to these surfaces. The observer may accelerate and rotate their tetrad, the spacetime may contain matter and a cosmological constant, and no symmetry is assumed. We impose regularity conditions at the vertex of each light cone and along the polar axis, including a general axial condition that allows for shear. Given a suitable prescription for the stresses of matter, the Einstein field equations can then be integrated hierarchically from data on a single light cone, and the regularity at the vertex fixes all constants of integration. The solution reveals a universal property: to first order in the observer's proper time, the evolution of the metric and of the four-momentum crossing the light cones is independent of the specific behavior of matter. As an application, we derive, for an observer immersed in matter, an analogue of the integral mass formula of black hole mechanics.
Read the full abstract

Differential Geometry

A Poincare-Einstein Liouville Theorem and the Yamabe Invariant of Lens Spaces

Schoen conjectured that the Yamabe invariant of every lens space is achieved by the round metric. A proof of this conjecture is presented for all lens spaces, together with rigidity: equality is only achieved for the round class. The key step is to show that every $\mathbb Z_p$-invariant Yamabe-positive conformal class on $S^3$ admits a $\mathbb Z_p$-equivariant Poincaré-Einstein filling. The distance function from a fixed point of the filling then gives an invariant test function with Yamabe quotient at most $\sigma(S^3)$, which descends to the lens space. The proof of the main theorem was generated by AI (ChatGPT Astra, September 2026); this is mainly a learning note of the AI generated proof.
Read the full abstract

All authors Liam Mazurowski, Tongrui Wang, Xuan Yao, Xin Zhou

Physics and Society

A Jump-Diffusion Model of Home Price Dynamics and Neighborhood Change

We develop a dynamical model that links the fluctuations of individual home prices to changes in the full distribution of home prices in an urban area. We model the price of an individual home as a jump-diffusion stochastic differential equation. To capture the dynamic and hyper-local economic conditions that drive the residential real estate market, we allow all parameters to depend on both time and neighborhood. In the mean-field limit, we derive the associated Fokker-Planck partial integro-differential equation (PIDE) governing the evolution of the home price distribution in a neighborhood. We quantify spatial heterogeneity within a city by using the Wasserstein-2 distance to compare home price distributions across the city's neighborhoods. We validate our model using data from Boston, Massachusetts as a case study.
Read the full abstract

All authors Ashley Peake, Adam Gilbert-Diamond, Jordan T. Kemp, Anette E. Hosoi

General Relativity and Quantum Cosmology

Weak-field analytical solutions for f(R,L_matt) gravity in a static and spherically symmetric space-time for perfect fluids

We study the weak-field regime of $f(R,\lmatt)$ gravity for a static and spherically symmetric configuration described by a perfect fluid, whose energy density is perturbed about a constant reference value. By means of a weak-field limit expansion about a Minkowski geometric background and under a quasi-Newtonian approximation, we derive and analytically solve the coupled system that governs the gravitational perturbations and the matter variables. The solution contains the long-range Newtonian contribution and Yukawa-type corrections in the temporal and radial components of the metric. In the matter sector, the energy density, the internal energy, and the pressure acquire radial contributions, with a non-zero constant background rest-mass energy density and vanishing background internal energy and pressure. The results show how the non-minimal coupling between curvature and matter jointly modifies the geometry and the fluid in the considered regime.
Read the full abstract

Computer Vision

MetricPose: Flexible 3D Human Pose Estimation with Depth and Geometric Cues

In this paper, we present a 3D human pose estimator that predicts metrically accurate human joint positions in the camera coordinate frame. This task requires recovering the configuration of the body joints, scale, and metric camera-local placement. In monocular images, body size and camera distance are coupled through perspective projection and therefore cannot be determined uniquely from image alone. This ambiguity can be reduced by combining three complementary sources of information: learned RGB priors, direct depth measurements, and geometric constraints (e.g., ground contact) from the environment. We extend a pretrained full-body estimator with optional metric depth and geometry-processing modules that improve camera-distance estimation and depth-guided body refinement while preserving RGB-only operation. The resulting unified model supports RGB, RGB-Ground, and RGB-D using shared learned parameters. Unlike pelvis-relative error, camera-local error also measures placement relative to the camera. Compared with the existing foundation model SAM 3D Body, our methods with RGB, RGB-Ground, and RGB-D modes reduce camera-local joint-position error by 12.16%, 31.00% and 54.61% respectively.
Read the full abstract

All authors Shifan Zhu, Sabad Modi, Hang Yang, Arushi Agrawal, Steven Gao, Meghan Huber, Nathan Wycoff, Donghyun Kim

Instrumentation and Methods for Astrophysics

System noise temperature measurements of a 19-element phased array feed

We present multi-epoch system noise temperature measurements of a 19 element hexagonal room-temperature phased array feed (PAF) using a modified Y-factor method. A Python-based pipeline with iterative channel calibration and maxSNR beam forming is developed to do data reduction, and a covariance-based decomposition scheme is proposed to extract mutual coupling noise temperature Tmc. Long-term observations from March to June 2026 capture temporal receiver noise drift mainly induced by LNA temperature variations. Within the 1.37-1.47 GHz operating band, the measured beam equivalent system noise temperature of the PAF is about 73-89 K. The measured Tmc baseline ranges from about 2 K to 8 K, while under thermal loading, the resonant peak of Tmc can increase to more than 20 K. The baseline of Tmc is consistent with the S-parameter mea surements, which show the Tmc of the full array ranges less than 10.7 K. The baseline fluctuation of Tmc most likely arises from the superposition of thermal and non-thermal components of the spatial electromagnetic fields. A detachable stainless steel shield is deployed to quantify sidelobe spillover noise, and comparative tests show unshielded en vironmental interference increases receiver noise temperature by over 60 K. The proposed measurements and data processing framework provides a practical reference for noise per formance characterization of ground-based radio telescope PAF prototypes.
Read the full abstract

All authors Shenghua Yu, Dejian Ding, Hongwei Xi, Chengjin Jin, Lei Yu, Bo Peng, Zhengwu Wang, Yan Zhu, Jianbin Li, Lijia Liu, Xinyu Zhu, Junjie Zhu, Jun Wang, Xiang Zhang, Chao Hu, Long Jiang, Xingyu Chen, Yaowei Li, Wenjie Wang, Zunli Yuan, Jiangming Chen, Hao Li, Zhaosheng Li, Xuetong Liu, Chengye Li, Zhiyong Wang, Dangen Xie, Hao Shang

General Relativity and Quantum Cosmology

Equatorial Periodic Orbits and Gravitational Wave Signatures in Weyl Geometric Gravity Black Holes

Weyl geometric gravity theory is a gravitational theory with Lagrangian density constructed from the square of the Weyl curvature scalar and the strength of the Weyl vector. The theory can be reformulated as a scalar - vector - tensor theory in the Riemann geometry by introducing an auxiliary scalar field, in the presence of a non - minimal coupling between the Ricci scalar and the scalar field. Assuming that the Weyl vector has only a radial component, an exact spherically symmetric black hole type solution of the field equations in the vacuum can be obtained. The solution depends on three arbitrary integration constants. As compared to the case of the Schwarzschild black hole, the Weyl geometric gravity black hole solution contains two new terms, linear and quadratic, in the radial coordinate. It is the main goal of the present work to investigate astrophysical properties of the Weyl geometric black holes by considering the behavior of the timelike geodesics, which allows to explore how the geometric nonmetricity influences the strong-field dynamics of massive test particles as well as the topological structure of the periodic orbits. The resulting gravitational wave emission in extreme mass-ratio inspiral (EMRI) systems is also considered. Our results show that the presence of nonmetricity in Weyl geometric gravity leaves distinct and measurable signatures on both the orbital topology and the emitted gravitational waveforms of compact EMRI systems. These unique temporal and amplitude modulations provide a promising observational avenue for testing the presence of nonmetricity on cosmic scales.
Read the full abstract

Materials Science

Mixed-dimensional topological charge compensation beyond Weyl-node pairing

The Nielsen-Ninomiya no-go theorem requires the total topological charge ($C$) in a periodic Brillouin zone to vanish, conventionally leading to opposite-chirality pairs of Weyl points (WPs). A charged nodal surface (NS) provides an alternative mechanism for compensating a same-chirality Weyl pair, yet the general symmetry constraints governing this configuration and its electronic realization remain unexplored. Here, by an exhaustive analysis of all 1651 magnetic space groups (MSGs), with and without spin-orbit coupling, we identify only 13 type-II MSGs that permit the realization of a minimal unpaired mixed-dimensional configuration comprising a single $|C|=2$ NS and two same-chirality $|C|=1$ WPs. Guided by this symmetry framework, we identify the chiral boron allotrope THC-B$_{30}$ as the first electronic realization of this unconventional topology. First-principles calculations reveal only two WPs with $C=-1$ and one NS with $C=+2$, satisfying global topological charge neutrality and exhibiting unconventional Fermi arcs connecting the WP and NS projections. Moreover, strain-induced topological phase transitions redistribute topological charge among distinct band-degeneracy manifolds. Our findings establish a symmetry-guided route to realizing same-chirality Weyl pairs through mixed-dimensional charge compensation beyond conventional Weyl-node pairing.
Read the full abstract

All authors Xiao-Jing Gao, Ke-Xin Pang, Yun-Yun Bai, Yanfeng Ge, Yan Gao

Multiagent Systems

AIM: Adaptive Interaction in Multi-Agent Debate for Multimodal LLM Inference

Multi-Agent Debate (MAD) improves reasoning on multimodal tasks, but its accuracy and token efficiency depend on how agents interact during debate. A largely overlooked design choice is which interaction modalities agents use to exchange evidence. Existing methods rely on fixed text- or graph-based interaction, which fails to convey fine-grained multimodal evidence and thus degrades accuracy while inflating token cost due to verbose descriptions. To address this, we propose Adaptive Interaction in Multi-Agent Debate (AIM), an adaptive routing framework that dynamically selects the optimal modality combination for inter-agent interaction in multimodal MAD. Beyond text and graph, AIM introduces task-dependent modality interaction, where agents exchange targeted regions of interest from the task-specific multimodal input, such as image regions, audio segments, or spatio-temporal video clips. To determine the best combination for each instance, AIM first generates a structured baseline response and extracts interpretable routing features that capture which evidence the instance requires, how confident the baseline agent is, and how much new evidence each modality would add beyond the baseline response. A lightweight router then uses these routing features to select a modality combination that avoids both under-fusion (i.e., agents miss decisive evidence) and over-fusion (i.e., redundant modalities inflate token cost) risks. Across eight benchmarks spanning three multimodal task families, AIM achieves the highest accuracy. Combining modalities improves accuracy by up to 6.4% over the best single-modality baseline, and adapting the combination per instance adds up to 4.2% over the most accurate fixed combination of each benchmark with up to 53.4% fewer tokens.
Read the full abstract

Networking

Age of Information for Power-Saving Devices with DRX Mechanism

Reducing the power consumption of User Equipment (UE) has always been a crucial issue in communication systems. As a result, Discontinuous reception (DRX) is introduced in today's New Radio (NR) networks as a power-saving strategy. In addition, the variety of low latency applications in 6G makes the issue of maintaining data freshness increasingly important. This paper aims to study the relationship between freshness and power-saving. To capture the information freshness, we use Age of Information (AoI) as the performance metric. To our best knowledge, unlike previous studies focusing on power saving and delay trade-offs, this is the first study considering the impact of DRX on AoI. We solve the closed-form solution of the peak average AoI with the DRX mechanism. Moreover, we propose a parameters adjustment strategy. Based on the simulation validation, our proposed strategy can enable the UE to keep the AoI close to the theoretical lower bound AoI while achieving high power-saving efficiency.
Read the full abstract

General Relativity and Quantum Cosmology

Landauer bound in a regular Bardeen black hole: extremality and evaporation

Does the vanishing Hawking temperature of an extremal black hole imply that information can be erased without energy transfer? We investigate this question in the regular Bardeen geometry using a two-level quantum system, representing a quantum memory, coupled to a scalar bosonic field confined within an exterior cavity. The Hawking temperature tends to zero at extremality, as does the corresponding thermal Landauer scale. The field in an exterior cavity behaves differently, provided both walls remain away from the horizon, its lowest excitation energy stays finite. For a maximally mixed memory coupled to a reservoir initially in its unique ground state, we find $Q_\infty\geq\Delta_{R,\infty}(1/2-\epsilon)$, with $\Delta_{R,\infty}$ the reservoir gap and $\epsilon$ the reset error. For a resonant single-mode interaction, we give a reset protocol that attains the bound with no work supplied by the external control. In this example, the memory itself supplies the energy gained by the field. In the multimode calculation, counter-rotating terms are retained. Keeping the pulse parameters unchanged, we find a larger residual reset error closer to extremality. For reservoirs prepared thermally, we evaluate the Reeb–Wolf balance and find that the relative-entropy contribution, expressed in energy units, can have a finite low-temperature limit. These numerical preparations are externally imposed and remain hotter than the Hawking temperature. For comparison, an unconfined scalar field has near-extremal luminosity $P_\infty\propto y^4$, with $y=r_H/|g|-\sqrt2$. Assuming a particular quasistatic energy-loss prescription gives $y\propto t^{-1/2}$; this does not establish the existence of a stable remnant. For this particular reset operation, vanishing Hawking temperature therefore does not force the transferred heat to vanish.
Read the full abstract

All authors J. Andrade, J. O. Cedeño, Jiří Svozilík, Tayyab Naseer, Faisal Javed, M. Zeeshan Gul

Distributed Computing

More with Less: Efficient Parallel Scientific Code Translation via an Agent-Skill Workflow

Translating parallel scientific codes between HPC programming frameworks is a critical but labor-intensive task. Existing LLM-based approaches rely on fixed sequential pipelines with undifferentiated repair strategies and correctness criteria that do not match scientific computing practice. We present an agent-skill framework that separates translation, compilation diagnosis, and execution validation into three autonomous agents coordinated by a verdict-driven orchestrator, routing each failure to a targeted repair skill rather than retrying blindly. An automated output check drives the self-correction loop, complemented by an offline LLM-based verification that exposes false positives escaping compiler and runtime checks. We evaluate the framework on 30 HeCBench kernels translated bidirectionally among CUDA, OpenMP, and SYCL with three open-source LLMs. Our framework attains the higher verified correctness rate in all 18 model and direction combinations over a state-of-the-art baseline while consuming up to 52.4% fewer tokens, remaining consistent across models and directions without architectural modification.
Read the full abstract

Artificial Intelligence

Flex-RSI: Benchmarking Recursive Self-Improvement of Frontier Agents

Frontier language models are increasingly deployed as agents, and how well an agent performs depends not only on its model but also on its harness: the prompts, tools, memory, and control loop through which the model acts. When an agent is applied to a new problem, a natural question is whether it can use the feedback of the task to improve the way it works, and not merely its answer. Most existing benchmarks keep the agent's harness fixed and score the solution it produces, or target large-scale model training. We introduce Flex-RSI, a benchmark for the recursive self-improvement of agents on vision and robotics tasks under a fixed budget. In each round, an executor improves the task solution, which is kept only if it scores higher on held-out validation data, and an improver then revises the harness, including the executor and itself. Every run starts from the same general-purpose harness, and the final solution is scored on hidden test data. Evaluating closed-source and open-weight models on tasks spanning perception, reasoning, robot action, 3D reconstruction, visual creation, and infrastructure, we find that self-improvement helps most when the agent builds its own solution, that feedback recording only success or failure holds back open-weight models, and that open-weight models fall furthest behind where outputs must pass strict checks of physical stability or exact correctness. Paired runs further show that editing the harness itself contributes to the gains.
Read the full abstract

All authors Jingdong Zhang, Jionghao Wang, Zhengming Yu, Yancheng Cai, Wei Chow, Chris Choy, Zhiyang Dou, Hang Hua, Jiachen Li, Zhixuan Liu, Chen Wang, Songlin Wei, Siyuan Yang, Yu Yuan, Xiaohang Zhan, Yucheng Zhang, Shuhong Zheng, Xin Li, Lingjie Liu, Wojciech Matusik, Zhengzhong Tu, Wenping Wang