Peer Revenue Now accepting positions

Papers

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

All of arXiv

Monday 9 November

Signal Processing

Integrating Deep PPG Representations and Digital Biomarkers for Aortic Stenosis Detection with Peri-TAVI Assessment

Access to echocardiography is often limited, delaying recognition of progressive aortic stenosis (AS). A brief, low-cost physiological measurement obtained during routine care could help identify patients who may warrant earlier imaging. Here, we present PulseAS, integrating deep photoplethysmography (PPG) representations, waveform biomarkers and clinical information for high-grade AS screening. Three complementary analyses assessed (1) validation and transfer across cohorts differing in geography and ethnic composition, (2) score-based enrichment of AS and pre-transcatheter aortic valve implantation (TAVI) cases, and (3) preoperative and postoperative pulse changes. Using 15-s recordings, separate patient-level cross-validation in EDU (n = 1,209) and MIMIC-III (n = 2,970) yielded AUROCs of 0.818 (PulseAS-XGB) and 0.838 (PulseAS-L2), respectively. In external validation from MIMIC-III to EDU, PulseAS-XGB achieved an AUROC of 0.716, improving to 0.813 after local adaptation. The highest-scoring 10% captured 45.2% of all high-grade AS cases in EDU and 41.3% in MIMIC-III, and 56.9% of pre-TAVI cases in EDU. Notably, in the EDU TAVI subgroup, PPG waveforms showed earlier systolic peaks and narrower systolic contours following intervention. These findings support clinical-PPG risk stratification in new populations and warrant prospective evaluation of echocardiographic prioritisation.
Read the full abstract

All authors Haowan Wang, Cheng Ding, Donglin Xie, Huizi Yu, Guangkun Nie, Shenda Hong, Bryan P Yan, Kent Chak-yu So

Combinatorics

A Cameron-Goethals-Seidel-Shult theorem for marked graphs

The classical Cameron-Goethals-Seidel-Shult theorem characterizes graphs with least eigenvalue at least $-2$ in terms of generalized line graphs, exceptional graphs, and root systems. In this paper, we establish a marked graph extension of this theorem. For a graph $G$ on $n$ vertices and a binary vector $\mathbf{b}\in\{0,1\}^n$, we completely characterize all marked graphs $\mathbf{G}=(G,\mathbf{b})$ such that $\lambda_{\min}(A(G)+\mathbf{b}\mathbf{b}^\top-J_n)\ge -2$, where $J_n$ denotes the all-ones matrix of order $n$. Using representations by finite root systems, we show that every irreducible marked graph satisfying the above inequality is represented by either the root system $B_n$ or the exceptional root system $E_8$. This leads to a complete structural characterization in terms of irreducible marked graphs, namely induced subgraphs of extended generalized line graphs, the marked graph $(2K_1,\mathbf{0})$, and exceptional graphs with all vertices marked. In the special case $\mathbf{b}=\mathbf{j}$, our characterization reduces to the original Cameron-Goethals-Seidel-Shult theorem. The root system $B_n$ and extended generalized line graphs play the roles occupied by $D_n$ and generalized line graphs in the classical theory. We further apply the characterization to graphs whose second largest eigenvalue is at most $1$, showing that every $-2$-markable graph satisfies this spectral condition.
Read the full abstract

All authors Chaohui Chen, Muhuo Liu, Haiying Shan, Zoran Stanić

Robotics

Action Representation for Spatial Generalization in Robot World Models

Spatial generalization in robot world models depends on whether action representations capture the physical relationships that remain valid across changing spatial configurations. Yet different coordinates can describe the same interaction, whereas similar motions can produce different outcomes when local physical conditions change. How these distinctions affect learned prediction remains insufficiently understood. This study evaluates matched action representations defined in world, object, and surface coordinate frames using a controlled robotic peeling benchmark. Global pose and contact relocation experiments reveal a cost associated with limited training data when equivalent interactions receive different spatial descriptions. Removing information about peel thickness creates an irreducible ambiguity because identical predictive inputs can correspond to different material removal outcomes. Coordinates defined relative to the surface improve consequence prediction and counterfactual action selection under accurate local observations, but learned predictors remain vulnerable to uncertainty in the coordinate frame. Joint centering of the surface and tool removes sensitivity to translation of their common origin, whereas incorrect contact geometry and actual execution deviations introduce different failure mechanisms. These results show that effective spatial generalization depends on preserving the relationships between actions and states that determine physical consequences, rather than simply minimizing coordinate variation. This perspective provides a principled basis for designing robot world models that reuse physical knowledge across spatial configurations while retaining the distinctions essential to prediction and planning.
Read the full abstract

Robotics

Fade to... Red? Scheduling Coloured Noise Exploration Supports Efficient Robot Learning with Reduced Movement Jerk

Exploration in deep reinforcement learning (RL) for continuous control is commonly driven by temporally uncorrelated Gaussian (white) action noise, which leads to jerky motions with locally constrained exploration. In contrast, temporally correlated coloured noise, particularly pink noise, has been shown to support more efficient exploration in RL. Motivated by these findings, we inquire whether scheduling changes to the noise colour over the course of training can outperform white noise for continuous control robotics. We perform a systematic study of 49 linear schedules between white and red noise, together with generalised state-dependent exploration (gSDE), on four sparse-reward manipulation tasks with a Fetch arm and a Shadow Dexterous Hand. We find that the noise colour at the end of training does not affect learning efficiency or end performance, whereas fading towards red noise halves the jerk of exploration in every task. Balancing learning efficiency, end performance and smoothness, red noise emerges as the preferred strategy: it learns as efficiently as pink noise with almost 2-fold less movement jerk. We therefore recommend exploration schedulers that fade towards red noise as the default for off-policy robot learning.
Read the full abstract

2611.11868 · 9 Nov · 10 pages, 4 figures, 1 table. Accepted at the ACRA 2026 conference. Cite as: "López, F. M., and Cruz, F. (2026). Fade to... Red? Scheduling Coloured Noise Exploration Supports Efficient Robot Learning with Reduced Movement Jerk. In Proceedings of the Australasian Conference on Robotics and Automation (ACRA). Australian Robotics and Automation Association (ARAA)."

Physics and Society

Competing Influence in LLM Populations: Budget Timing and Baseline Drift

When competing campaigns target a population of large language model (LLM) agents, observed opinion changes may reflect both campaign influence and changes that would occur without intervention. We study how the timing of campaign expenditure and the population's own evolution shape outcomes in a network of persona-conditioned LLM agents that exchange and revise opinions in natural language. A proponent and a challenger are tasked with generating opposing messages under equal delivery budgets, while a separate stance probe measures opinions without influencing the simulation. Across eight claims, we compare single-campaign and competitive conditions against matched no-campaign runs, and evaluate all nine combinations of early, sustained, and late spending on a reference topic. Baseline stance drift varies substantially across claims and, on some, accounts for a sizeable fraction of apparent campaign effects. After accounting for this drift, the direction of observed campaign-associated changes aligns more closely with the claim's epistemic standing than with the campaigner's assigned role. Uninfluenced populations also consistently lose stance diversity, demonstrating that directional opinion change and convergence are distinct phenomena. In networks with two opinion-segregated communities, campaigns primarily reduce differences between communities rather than diversity within them. LLM campaign agents do not always argue for their assigned positions, particularly when these conflict with established evidence. Finally, campaign outcomes depend on the opponent's expenditure schedule: unlike a matched numerical bounded-confidence population receiving interventions at the same times and locations, the LLM population exhibits a payoff structure with no pure-strategy equilibrium.
Read the full abstract

Distributed Computing

DeFSM: Robust and Efficient Decentralization of Federated Submodular Maximization

Submodular maximization seeks a small subset best representing a large dataset. In its federated setting, the server is a single point of failure and some clients may be Byzantine. Removing the server raises a trilemma of decentraliza- tion, efficiency, and security. We propose DeFSM, a Byzantine-robust and communication-efficient architecture that replaces the server with a lightweight smart contract and a pool of independent, consensus-free agents. Its aggregation rule, SafeGM, builds on two notions from combinatorial geometry: Tukey depth and safe points (the points lying in the convex hull of the good points wherever the bad points are). SafeGM makes the robust decision on 2D Gaussian random projections, which the contract can compute efficiently, and lifts it back as a convex combination of actual updates. Since lifting loses information, it uses the projected geometric median only when its depth certifies it, and otherwise a Tukey median. The share of bad updates is then bounded only by a count: how many there are, not what they contain. We prove, and confirm empirically, that this bound controls the solution quality. On CIFAR10 and PathMNIST, with up to 30% Byzantine clients and agents, DeFSM never collapses and retains on average about 91% of the quality of the ideal reference, the geometric median at a trusted central server. The geometric median decentralized in its best setting would retain about 96%, but the contract cannot compute it, and so each agent would have to download six times more data, or each client twice as much.
Read the full abstract

Group Theory

Schur multiplier exponents of finite $p$-groups: bounds and critical-class phenomena

The Schur exponent conjecture is a long-standing conjecture in group theory, which asserts that $\operatorname{exp}(\mathrm{M}(G))\mid\operatorname{exp}(G)$ for every finite group $G$, where $\mathrm{M}(G)$ denotes the Schur multiplier of $G$. In this article, we prove that the Schur exponent conjecture holds for finite $p$-groups $G$ of nilpotency class at most $2p-2$ when $p$ is an odd prime. Our approach is to establish upper bounds on the exponent of $\gamma_{s+1}(G)$ for $s\in\mathbb{N}$. Using this approach, we further establish exponent bounds for Schur multipliers, nonabelian exterior squares, and tensor squares of $G$ of arbitrary nilpotency class, improving some existing bounds. Moreover, we show that Moravec's Bogomolov multiplier exponent conjecture holds for finite $p$-groups of nilpotency class at most $2p-1$. Havas conjectured that, for every prime $p>3$, there exists a finite $p$-group of exponent $p$ and nilpotency class $2p-1$ for which the Schur exponent conjecture fails. We show that the Schur exponent conjecture still holds for finite $p$-groups of nilpotency class at most $2p-1$ generated by at most three elements. For four-generator groups of exponent $p$ and class at most $2p-1$, we prove that $\mathrm{M}(G)\cong C_{p^2}^{\,b}\times C_p^{\,a}$ for some $a\ge0$ and $b\in\{0,1\}$.
Read the full abstract

Theoretical Economics

Separable social welfare evaluation without completeness

In this paper, we study incomplete social evaluation of utility vectors to capture situations in which there is disagreement in society about the degree of inequality aversion. We characterize multi-prioritarian criteria using efficiency, equity, impartiality, separability, and a condition called continuous divisible completion condition. We also provide uniqueness results and investigate comparative inequality aversion.
Read the full abstract

Machine Learning

Can Hindsight Critics Guide Delayed Actors?

Sensory delays create temporal misalignments between the state of the environment and the observations available to an agent, turning Markovian control problems into partially observable ones. In this work, we inquire whether such misalignments can be mitigated during training by using hindsight to recover aligned state-action transitions. We introduce Hindsight Critics, an asymmetric extension of the Soft Actor-Critic for reinforcement learning under partial observability. The critics are trained on hindsight-aligned transitions whereas the actor learns from the resulting privileged value estimates despite only having access to delayed observations. Comparisons against delay-aware and unaware baselines on the DelayBench suite show that Hindsight Critics can partially compensate for observation delays without requiring modifications to the agent's observations during deployment.
Read the full abstract

2611.11863 · 9 Nov · 10 pages, 1 figure, 3 tables. Accepted for publication at the AJCAI2026 conference. Cite as: "López, F. M., & Cruz, F. (2026). Can Hindsight Critics Guide Delayed Actors?. In Australasian Joint Conference on Artificial Intelligence (pp. 1-10). Cham: Springer International Publishing."

Machine Learning

Are UFOs Real? Introducing Unified Factored Organizers

Implicit in much research in deep learning is the assumption that if a network achieves low loss, then it must have learned a good representation of its training data. Recent work by Kumar et al. (2025, arXiv:2505.11581) questions this assumption, arguing that networks trained with conventional gradient descent achieve low loss despite learning fractured entangled representations (FER), leaving related concepts disconnected and unrelated concepts entangled inappropriately. Kumar et al. further show that dramatically better representations are possible: a network evolved through an open-ended, human-in-the-loop process learned unified factored representations (UFR) that capture the regularities of the domain far better than a counterpart with nearly identical outputs trained with conventional gradient descent. Inspired by UFR, this paper introduces unified factored organizers (UFOs), networks trained to model regularities in the representations of other networks. A network can then be trained both to achieve low loss on its task objective and to align its representations with the model learned by the UFO. Networks trained with UFOs are shown to learn qualitatively superior representations in an image-regression domain where representations are easily interpretable. To provide quantitative evidence for the impact of UFOs on representation learning, a transformer is trained with and without UFO regularization to grok a modular arithmetic problem set. When combined with weight decay, UFOs can reduce the number of transformer training steps needed to grok by more than 10x relative to weight decay alone. Without weight decay, UFOs can reduce the number of transformer training steps needed to grok by more than 100x relative to an unregularized baseline. If UFOs are real, they suggest a new paradigm where loss functions themselves are trained to learn an effective model of good representation.
Read the full abstract

All authors Thomas X. Sheppard, Alon Albalak, Joel Lehman, Kenneth O. Stanley

Cryptography and Security

Scaling Adaptive Prompt Injection Attacks with Agent Swarms

Prompt injections have emerged as one of the top security risks faced by AI agents today. Early work in prompt injection attacks looked at simple tool-call agents and static attack payloads, and were saturated since mid-2025. As of Fall 2026, human red-teaming may be nearing saturation, especially on the Gemini family of models. In this work, we argue that despite this apparent plateau, prompt injection attacks are still possible and continued effort into prompt injection defense is needed. We design a new synthetic evaluation dataset which uses random non-malicious actions as adversarial goals. Such dataset design helps us to evaluate prompt injection robustness as a model ability to respect instruction hierarchy, and disentangle this ability from model knowledge of security practices. We show that an agentic attacker of increased capabilities can reliably trick model to follow instructions in tool output in our evaluation, even for models previously considered robust on public prompt injection benchmarks such as GraySwan Arena. This shows that current frontier defenses may require proactive, automated red-teaming to remain resilient to increasingly capable adversaries.
Read the full abstract

All authors Alexey Kurakin, Vishal Karande, Alex Bie, Chongyang Shi, Shuang Song, Saeed Mahloujifar, Abdullatif Köksal, Jamie Hayes, Juliette Pluto, Yinghao Sun, Kamalika Chaudhuri

Computational Physics

GPU acceleration of the numerical-atomic-orbital DFT code OpenMX with GEMMul8: performance and numerical agreement on consumer and HPC GPUs

We integrate GEMMul8, a library that emulates FP64 matrix multiplication on INT8 Tensor Cores, into the GPU version of the numerical-atomic-orbital first-principles code OpenMX 4.0. The GEMMul8 configuration is compared against the cuBLAS FP64 configuration within each of two environments, a consumer GeForce RTX 5080 and an HPC-oriented H100 PCIe, for the four computational paths formed by collinear (CL) or noncollinear (NCL) spin and band or $\Gamma$-point-cluster solvers. In fixed-work measurements that completed 25 SCF iterations, the diagonalization section of a 384-atom CL band calculation on the RTX 5080 is accelerated by a factor of 2.1 and the total run time by 1.8; a 512-atom CL cluster calculation, which uses only real matrix products in contrast to the CL band calculation, still gains a factor of 1.5 in the diagonalization section. On the H100, GEMMul8 provides little additional acceleration of the diagonalization section. In an accuracy test on a 216-atom Si system in which one atom was displaced by 0.05 angstrom, all 24 configurations were converged using an SCF convergence criterion of $10^{-8}$; every configuration converged at the same iteration within each path, and the differences between GEMMul8 and cuBLAS are at most $2\times10^{-12}$ Ha in the total energy and $7\times10^{-12}$ Ha/Bohr in force components.
Read the full abstract

Cryptography and Security

Doubly-Efficient Private Information Retrieval from Modular Forms

We present new information-theoretic protocols for $s$-server, doubly-efficient private information retrieval (PIR) in sublinear server time and one extra bit of space on the servers, for every constant $s \geq 2$. On an $n$-bit database, each of the $s$ PIR servers first builds and stores a preprocessing data structure consisting of $n+1$ bits (which is unique to each server, and after which the server may discard the original database). Then, each server answers a PIR query in time $T_s(n)$, by reading at most $T_s$ bits of its data structure and sending them to the user. In the two server setting ($s=2$), our PIR protocols achieve $T_2 = n^{0.82}$, matching the server time of the state-of-the-art schemes of Henzinger and Ragavan (EUROCRYPT 2026) while shrinking their storage from $O(n\sqrt{\log n})$ to $n+1$ bits. With $s \geq 3$ servers, our PIR protocols retain the $n+1$ bits of storage and additionally improve on the server time of all prior information-theoretic $s$-server schemes with $n^{1+o(1)}$ storage: for example, achieving $T_3 = n^{0.74}$ and $T_{10} = n^{0.61}$ (previously $n^{0.79}$ and $n^{0.75}$). Asymptotically, as the server count $s$ grows, we shrink the number of servers required to achieve a given server-work exponent from $s^{3.9}$ to $s$. Our results draw inspiration from matching-vector techniques (Efremenko, STOC 2009), which we generalize to handle degree-two computations in the exponent. Using this generalization, we build doubly-efficient PIR from the new intermediate abstraction of Fourier-sparse cone-delta functions. Then, we construct such Fourier-sparse cone-delta functions via a connection to modular forms and complex analysis: if a Fourier-sparse cone-delta function does not exist, then there exists a modular form which has too many roots, contradicting the valence formula.
Read the full abstract

Quantum Physics

Rotational Encoding of Objective Functions for Quantum Optimization

In this paper, we introduce the Angular Rotational Counter (ARC), a novel approach to represent objective functions that can be decomposed as a weighted sum of Boolean functions. ARC encodes the objective as an accumulated sum of quantum rotations onto a single-qubit register instead of a multi-qubit register used in much of the previous literature. We mathematically establish and experimentally validate that these rotations sum to the objective value for all possible solutions and derive the relationship between the resulting measurement probabilities and the objective. This design was applied to construct a quantum oracle for the state assignment problem of finite automata and state machines using a weighted Hamming distance cost function, which is a special case of the quadratic assignment problem. We evaluate this oracle design on 26 benchmarks and compare it with corresponding Boolean-based oracles on four cost metrics. Across these benchmarks, the oracle with ARC produces a mean reduction in the qubit count of 20.8%, gate count of 89.3%, circuit depth of 92.4%, and quantum cost of 90.6%. These results demonstrate substantial reductions in circuit cost for the evaluated benchmarks and establish rotational encoding as a promising approach for representing objective functions in optimization based on quantum search algorithms.
Read the full abstract

All authors Sonia Yang, Ali Al-Bayaty, Shanyan Chen, Marek Perkowski

Distributed Computing

OverlapAdvisor: Modeling and Adapting Computation-Communication Overlap for Distributed MoE Serving

Mixture-of-Experts (MoE) inference with expert parallelism (EP) relies on All-to-All communication, making computation-communication overlap strategies critical for achieving high performance. However, no strategy is universally optimal. Across 103 production configurations, varying input lengths, and four hardware platforms, we observe that the fastest strategy shifts based on model shape, EP degree, input length, and hardware. Since no strategy dominates more than 80.6% of configurations, configuration-aware selection is necessary. In this work, we propose OverlapAdvisor, which models overlap stages using two granularities to capture computation-communication dependencies. OverlapAdvisor builds a unified performance model, accounting for deviations from peak compute, memory, and communication performance across workloads and hardware. This model enables launch-time, configuration-aware strategy selection within two serving engines. We conduct experiments on four MoE models (30B-235B) against three baselines, utilizing SGLang and vLLM. At the kernel level, OverlapAdvisor achieves 1.16x average and 1.59x maximum speedups, with a prediction MAPE of 3.3%-15.9%. In SGLang and vLLM, it improves end-to-end throughput up to 31.7%. Its granularity-guided optimizations further achieve median speedups of 1.10x for DeepEP-v1 and 1.22x for FlashMoE.
Read the full abstract

All authors Chutong Ding, Kaixuan Zhang, Luping Wang, Shiyou Qian, Jian Cao, Cheng Huang, Guodong Yang, Liping Zhang

Machine Learning

DeltaForge: A Post-hoc Mixture-of-Experts VLA for Multi-Domain Autonomous Driving

Fleets of driving vision-language-action (VLA) models are developed one domain at a time: each operator finetunes a shared base model on private data, producing specialized experts that are costly to consolidate into a single model. Weight merging loses accuracy as heterogeneous domains accumulate, and dense mixture-of-experts routing is prohibitively expensive to combine them. We present DeltaForge, to our knowledge the first post-hoc mixture-of-experts VLA for driving, which consolidates N independently finetuned per-domain models into a single multi-domain model from their weights and a small proxy set alone. When different operators finetune the same base model, their weight updates end up sharing a lot of structure. We exploit this by decomposing all the updates together using singular value decomposition (SVD), splitting each one into a shared part (the same across all domains) and a tiny domain-specific part. The router then only needs to work with those tiny domain-specific pieces, not the full expert weights, which is what keeps the approach efficient. Then two cooperating clip-level blind routers, one over the VLM reasoning experts and one over the flow-matching dynamics experts, are trained from soft labels on the proxy dataset. Across six diverse domains spanning urban, off-road, and highway driving on cars, trucks, and all-terrain vehicles (ATVs), DeltaForge approaches the accuracy of the per-domain experts and surpasses every merging, routing, centralized, and federated baseline at close to single-model parameter and compute cost. Adding a new domain requires only projecting its weight update onto the shared basis with no end-to-end retraining needed. The resulting model also generalizes to domain combinations it was never explicitly trained on.
Read the full abstract

Theoretical Economics

Distributionally Robust Rational Inattention

We study a rational inattention model in which a decision maker chooses a stochastic choice rule while facing ambiguity about the prior distribution over states. Ambiguity is modeled through a distributionally robust formulation in which the prior belongs to an uncertainty set defined by a general phi-divergence. Our main result establishes an exact Lagrangian dual representation of the resulting max-min problem as a finite-dimensional concave optimization problem. This characterization yields necessary and sufficient conditions for the emergence of robust consideration sets and a tractable one-dimensional dual representation for the Cressie-Read family of divergences. We further connect the model to robust risk measures: Average Value-at-Risk arises from a particular non-smooth divergence, while Kullback-Leibler ambiguity is linked through Lagrangian duality to the entropic-risk multiplier representation. We then develop two extensions. The first allows the attention-cost technology and prior ambiguity to be governed by distinct phi-divergences, while the second introduces a deeper layer of ambiguity over the decision maker's baseline prior. Both extensions preserve the finite-dimensional structure of the model and encompass standard Kullback-Leibler robust rational inattention specifications as special or corresponding multiplier cases.
Read the full abstract

Robotics

Decoupled Asynchronous System for Humanoids (DASH): Latency-Agnostic Scaling of Vision-Language-Action Models for Humanoid Manipulation

Vision-language-action (VLA) models pair a vision-language backbone with an action head, but in closed-loop control the backbone sits on the control-critical path: while it computes, the robot keeps moving, so each executed action is conditioned on an increasingly stale observation. A larger backbone therefore directly blocks reactivity, especially in humanoid dexterous manipulation. We present DASH, a Decoupled Asynchronous System for Humanoids that removes this coupling. DASH splits the policy into a slow vision-language model (System 2) and a fast flow-matching action head (System 1) that run as concurrent processes over the LCM communication protocol, so the control rate is set by the action head alone, independent of backbone size. The action head fuses stale semantics from a frozen backbone with fresh vision from its own ResNet encoder, desynchronized training exposes it to the staleness it meets at deployment, and an adaptive temporal aggregator aligns overlapping action chunks under variable inference latency. On an asynchronous RoboTwin benchmark that charges inference latency in simulated time, DASH holds 20 Hz control rate from 0.8B to 8B backbones while the synchronous rate falls from 9.7 to 4.1 Hz, with higher success at every size. On a real humanoid with dexterous hands, DASH achieves clearly higher pick-and-place success than the synchronous pipeline.
Read the full abstract

All authors Bingzhang Wang, Nathan Zhao, Jialeng Ni, Jiayin Zheng, Yin Zhang, Kehua Chen, Jian Yao

Solar and Stellar Astrophysics

The Dusty Circumstellar Structure of the Yellow Hypergiant IRC +10420: Constraints from the SED and Near-Infrared Imaging Polarimetry

The yellow hypergiant IRC +10420 is a post-red supergiant evolving toward higher temperatures. We model its circumstellar dust with the radiative transfer code RADMC-3D, including polarized scattering, constrained by the spectral energy distribution (SED) from 0.4 $\mu$m to 1.3 mm and by the K' (2.2 $\mu$m) imaging polarimetry of Shenoy et al. (2015), assuming a pole-on view. The model consists of an equatorial structure (a flared inner disk of 0.3 $\mu$m grains at 300-3000 au, a ring at 3000-16,000 au whose surface density peaks at $\approx 10^4$ au, and a thin mid-plane disk of 300 $\mu$m grains) embedded in a spherical envelope of 0.15 $\mu$m grains with $\rho \propto r^{-2.3}$. It reproduces the SED within 0.07 dex and the total- and polarized-intensity profiles; the fractional polarization peaks at 15% at 1.7 arcsec (observed: 14%). The rising surface density of the ring produces the polarized-intensity plateau, and 0.5 $\mu$m grains at its outer edge the decline beyond 2 arcsec; grains of $\gtrsim 0.7$ $\mu$m are excluded. A thin disk of large grains suggests a settled mid-plane. Stellar temperatures of 7000-8500 K fit equally well. Of the total extinction ($A_V \approx 6.9$ mag), 5 mag is interstellar. Most of the $7.4 \times 10^{-2}$ $M_\odot$ of dust is equatorial; the envelope implies a mass-loss rate of $\approx 10^{-5}$-$10^{-4}$ ($\psi$/100) $M_\odot$ yr$^{-1}$. Because equatorial dust redirects starlight poleward, the pole-on SED overestimates the luminosity by $\approx 1.5$; at the Gaia distance, $L_\star \approx 3.3 \times 10^5$ $L_\odot$. Subtracting a scaled stellar PSF removes most of the nebular light within $\approx 1$ arcsec.
Read the full abstract

Machine Learning

Identifiable but Hard to Estimate: Rotation in Equally Spaced Overcomplete ICA

How many observations are needed to estimate a common rotation in two-dimensional overcomplete independent component analysis? We study equally spaced mixing columns with known source count, common source law, and column norm. For a fixed non-Gaussian fourth-order Gram-Charlier source law, the rotation is identifiable modulo the column spacing. Yet the sample size for fixed relative angular accuracy grows exponentially with the source count. We characterize this sample size, up to absolute constants, by the reciprocal of the squared Hellinger distance between two rotations separated by half the column spacing. An explicit radially weighted phase estimator attains this scale uniformly over the rotation family, using the observation radius to prevent cancellation in the mean of an angular harmonic. We also determine the exact exponential decay rate of single-observation total variation for the same pair. This gives bounds on the sample size exponent; its exact value remains open. Finally, a noisy binary-source example shows how an exact zero line of the observed characteristic function can have an exponentially small normal derivative.
Read the full abstract

All authors Wenjun Xiao, Yinghao Fang, Hanxin Chen, Sihan Jia, Vince D Calhoun, Yuda BI

Analysis of PDEs

An operator theoretical approach to weak solvability of a porous medium equation with Neumann boundary condition

This paper is concerned with the Neumann problem for the porous medium equation, \[ u_t - \Delta a(u) = f \quad \mbox{in}\ \Omega \times (0, T), \qquad \nabla a(u) \cdot \nu =0 \quad \mbox{on}\ \partial \Omega \times (0, T), \qquad u(\cdot, 0) = u_0 \quad \mbox{in}\ \Omega \] in a smoothly bounded domain $\Omega \subset \mathbb{R}^N$ $(N \in \mathbb{N})$, where $T \in (0,\infty)$, $\nu$ is the outward unit normal vector to $\partial \Omega$, and where $f$ and $u_0$ are given. A typical example in mind is $a(u)=|u|^{p-2}u$ with $p \in [2, \infty)$. The main interest is weak solvability of the problem in the case that $u_0 \in H^1(\Omega)^*$, $f \in L^2(0,T; H^1(\Omega)^*)$. Under the condition that the mean value of $f(t)$ vanishes for almost all $t \in (0,T)$, existence of weak solutions is established by an application of abstract results from operator theory, for instance, properties of Gâteaux derivative. The key of the proof lies in the approximate equation \[ [(\varepsilon I - \Delta)^{-1} u_\varepsilon]_t + a(u_\varepsilon) = (\varepsilon I - \Delta)^{-1} f, \quad \varepsilon \in (0,1), \] where the map $\varepsilon I - \Delta$ from $H^1(\Omega)$ onto its dual space is regarded as the Riesz isomorphism of $H^1(\Omega)$ equipped with an appropriate topology.
Read the full abstract

Statistical Mechanics

Computing Classical Partition Functions via Cycle-Space Sampling

We present a method for computing thermodynamic properties of classical spin systems by reformulating the partition function as a sum over even subgraphs of the interaction graph. Starting from the high-temperature expansion, we show that the expansion coefficients are determined by the weight distribution of the null space of the binary incidence matrix over $\mathbb{F}_2$, and that this statement holds verbatim for multi-spin interactions once the graph is replaced by a hypergraph. We introduce a Monte Carlo scheme based on the $1/t$ Wang-Landau algorithm that performs a random walk directly in this null space. Because the weight distribution is temperature independent, a single estimate yields the zero-field partition function, and the thermodynamic quantities obtained from its temperature derivatives, at arbitrary temperature. We validate the approach against exact finite-size results on the two-dimensional triangular-lattice Ising model, apply it to the three- and four-dimensional hypercubic Ising models where no exact solution is available, and to the two- and three-dimensional Baxter-Wu models, whose three-spin interactions place them outside the reach of Pfaffian methods. For the two-body models we compare against conventional Wang-Landau sampling in energy space; for the Baxter-Wu models we validate against exact enumeration at the smallest size in each dimension.
Read the full abstract

Information Theory

Joint LSFC and Nakagami-m Estimation From Downlink SINR in GZF-Precoded CF-mMIMO

Accurate characterization of the instantaneous signal-to-interference-plus-noise ratio (SINR) is important for reliability-sensitive cell-free massive multiple-input multiple-output (CF-mMIMO) systems. This paper proposes a joint estimation method for link-specific large-scale fading coefficients (LSFCs) and a common Nakagami-$m$ shape parameter using power and downlink SINR measurements. A tractable SINR distribution is first derived under global zero-forcing (GZF) precoding, Nakagami-$m$ fading, and link-level fractional power control (FPC). Based on the derived probability density function (PDF), a model-driven neural network (NN) is developed to solve the resulting maximum-likelihood estimation (MLE) problem. Simulations show that the reconstructed SINR distributions obtained with the estimated parameters closely match the empirical distributions. With $2 \times 10^{5}$ samples, the proposed method achieves an LSFC root-mean-square error (RMSE) of approximately 1.0 dB and an $m$ RMSE of approximately 0.085.
Read the full abstract

All authors Longfei Liu, Yichen Guo, Lili Xu, Yihang Cheng, Tao Peng, Wenbo Wang

Artificial Intelligence

Causal Signal Credit Assignment for Verilog Code Generation via Counterfactual Hardware Verification

Large language models have improved Verilog generation, yet functional correctness remains difficult to optimize because simulation commonly provides a coarse module-level pass/fail outcome. Such feedback discards useful evidence from partially correct RTL modules and cannot determine whether a source statement is functionally responsible for an observed output behavior. We present a new method named Causal Signal Credit Assignment (CSCA), which is a counterfactual preference-optimization framework for Verilog generation. CSCA first performs continuous signal-level behavioral verification on sampled candidate modules. It then locates output-driving assignment spans and applies five interventions while holding the remaining module and testbench fixed. The resulting changes in signal-level errors are combined with behavioral agreement to form a continuous statement credit, which is mapped to response-token weights for strict weighted direct preference optimization. This enables CSCA to learn from locally supported statements without uniformly reinforcing an otherwise incorrect response. Built on Icarus Verilog and a CodeV-derived corpus, CSCA is evaluated on VerilogEval~2.0 and RTLLM~v2.0, including both code completion and specification-to-RTL tasks. With a Qwen2.5-Coder-7B-Instruct backbone, CSCA leads the evaluated 7B-scale models and remains comparable to the much larger GPT-5.4-mini on the reported Verilog-generation benchmarks.
Read the full abstract

Machine Learning

Severity-Controlled Prediction Sets for Medication Recommendation

Medication recommendation is a multi-label prediction task. However, existing prediction-set evaluation and calibration methods rely mainly on exact-overlap metrics and therefore treat all missed medications as equally severe, overlooking how closely the prediction set matches them within the medication hierarchy. To address this, we introduce Conformal Severity Control (CSC), a post-hoc framework for calibrating medication prediction sets under hierarchy-aware mismatch severity. CSC uses the Anatomical Therapeutic Chemical (ATC) hierarchy to quantify mismatch severity and controls patient-level severity without retraining the predictor, supporting both average severity and upper-quantile severity for the highest-severity mismatches. Experiments on MIMIC-IV show that CSC tracks user-specified severity tolerances and shifts prediction sets away from distant mismatches toward exact or close matches.
Read the full abstract

All authors Yu Gu, Zijun Yu, Chi-Kuang Yeh, Xinyu Wang, Ziyang Song

Artificial Intelligence

MOTIF: LLM-Defined Material Taxonomies for OOD Property Prediction via Invariant Feature Selection

Out-of-distribution (OOD) generalization remains a central challenge in materials property prediction, where models must extrapolate to novel chemistries beyond the training distribution. While large language models (LLMs) have demonstrated strong scientific reasoning, their direct use for quantitative prediction remains limited. To address this, we systematically evaluate multiple roles for LLM assistance, finding that only taxonomy-based grouping consistently improves OOD generalization over a strong kNN+GBT baseline. Building on this, we here propose MOTIF, a framework that leverages LLMs not as predictors, but as generators of candidate taxonomies that partition materials into chemically meaningful environments. These taxonomies are evaluated via training-only pseudo-OOD validation, and only those that improve generalization are retained. Invariant feature selection is then applied within each accepted taxonomy, followed by gated aggregation across taxonomies. We evaluate on 2D halide perovskites, where bandgap prediction must generalize across distinct organic spacer families. On this benchmark, MOTIF reduces mean OOD MAE from 0.122 to 0.111 eV. Under a frozen dev/test protocol, the deployable configuration further improves errors from 0.123 to 0.117 eV (electronic bandgap) and from 0.181 to 0.173 eV (optical bandgap). Beyond this primary task, we validate transferability across additional materials datasets, including 3D perovskite solar cell efficiencies and Matbench benchmarks with structure-based splits. Across domains, MOTIF consistently improves robustness under distribution shift, outperforming expert-defined groupings, rule-based heuristics, and clustering baselines, while remaining competitive with strong classical selection methods. These results demonstrate that LLMs are most effective not as predictors, but as tools for discovering structure in data.
Read the full abstract

All authors Gwangseon Jang, Sieun Yun, Dong Ho Lee, Kyong-Ha Lee, Chanuk Lim, Jin Woo Choi, Han Seul Kim

Signal Processing

Prop-WRF: Proposal-Guided Wireless Radiance Field for Scalable Multi-Frequency Outdoor Channel Frequency Response Prediction

Sparse-directional wireless radiance fields (WRFs) enable complex channel frequency response (CFR) prediction from receiver locations and multipath direction priors. However, in large-scale outdoor environments, long sampling intervals make it difficult to concentrate a limited sampling budget around valid virtual anchors (VAs), degrading prediction accuracy and training stability. This paper proposes Prop-WRF, a proposalnetwork-guided WRF for CFR prediction across multiple subcarriers. A lightweight proposal network learns VA occupancy densities through online distillation from the main network and iteratively guides resampling toward likely VA locations. The main network then predicts complex radiance over the refined intervals and synthesizes CFRs using an analytical propagation model. By introducing frequency encoding only into the radiance branch, a single model jointly learns shared VA geometry from observations across subcarriers while representing frequencydependent complex radiance. A stage-wise hybrid normalized loss and an interval-weight regularizer further strengthen weakchannel supervision and suppress spurious VA peaks. In addition, a multipath simultaneous localization and mapping (SLAM) front end refines receiver states and path associations to construct geometrically consistent training priors. Simulations show that Prop-WRF improves CFR prediction accuracy and training stability over conventional sparse-directional WRF baselines in larger environments.
Read the full abstract

All authors Yulin Fu, Jiancun Fan, Shiyu Zhai, Zhibo Duan, Jie Luo

Geometric Topology

Compact branch sets and injectivity of proper branched coverings

In this paper, we show that a proper continuous open discrete map from the open unit ball $B^n$ onto a domain in $\R^n$ with $n\geq5$ is a homeomorphism whenever its branch set is compact. This answers Vuorinen's question affirmatively in every dimension $n\ge5.$ In dimension four, we also show that, under the additional assumption of quasiregularity, the map is a quasiconformal homeomorphism. Together with the three-dimensional theorem of Kauranen, Luisto and Tengvall, these results answer the quasiregular question posed as Problem~7.66 in the collections of Barth–Brannan–Hayman and Hayman–Lingham in every dimension $n\ge3$.
Read the full abstract

Computer Vision

Aggregation-Induced Calibration Failure in Sum-Based Label Expansion for Multi-Label Facial Expression Recognition: A Diagnostic Study

Label-space expansion - predicting a larger joint space and summing back to the original labels - is a natural design for multi-label facial expression recognition, but its failure modes have not been characterised. We evaluate Condition-Aware Label Expansion (CALE) on a self-built dataset (30,240 training images, 3,460 multi-label test images), on seven backbones, and on the public RAF-DB and RAF-ML benchmarks. The central finding is that the aggregation rule, not the representation, decides whether the method works. With a fixed threshold of 0.5, the aggregate score spans 28x across five backbones at 10% of the training data (0.24 to 6.79) and collapses to a 1.9x range at full data, while threshold-free mAP stays within 1.28x over the same backbones, which localises the fault to the summation step. The mechanism is the uncalibrated aggregation rule, which adds the main-branch score to half the sum of the three non-main branches; we verify it by measuring main- and non-main-branch activations. Two plain seven-dimensional sigmoid baselines (BCE, ASL) reach mAP 0.5055 and 0.5175; ASL exceeds three of the five CALE configurations and falls within 0.03 of the best. On single-label FER the gain over a seven-dimensional baseline is nominally significant (paired t-test, p = 0.0436) but does not survive Benjamini-Hochberg correction (p_BH = 0.2055) or a Wilcoxon test (p = 0.1250). We conclude that sum-based label expansion requires threshold calibration at the deployment data scale, or a learned aggregation, before it can support multi-label decisions.
Read the full abstract

Robotics

A Benchmark and Layered Evaluation Protocol for Evidence-Grounded Shared Belief Updating in Connected Robotics

Connected physical systems fuse observations from robots, IoT sensors, and edge-AI modules into a shared account of the world, yet most evaluations score only the final answer or action and cannot tell whether a method attributed the evidence, updated its belief, handled contradictions, and justified its decision. We argue that shared reasoning in Connected Robotics should be evaluated as an evidence-to-belief trajectory, and introduce the Connected Robotics Evidence Graph (CREG) Benchmark to make that trajectory measurable. CREG comprises 350 deterministically generated scenarios across seven regimes (justified revision, belief preservation, relation association, contradiction onset and resolution, stale/delayed evidence, and source reliability), each exposing a normative posterior reference at every evidence prefix, validated by an independent log-odds checker and screened by a blinded expert audit. Nine representation conditions separate relation content from serialization, ordering, and metadata. Evaluating four open-weight instruction models and six transparent fusion baselines yields three findings. First, entity grounding accounts for a substantial part of the weak-model deficit. Second, deleting semantic relations costs 15.3 points of strict belief-state accuracy on relation-dependent scenarios (Holm-corrected p<0.001), but an information-matched structured list outperforms the full graph and shuffling the graph does not hurt: relation content matters where required, while graph serialization confers no consistent advantage. Third, belief accuracy is strongly non-monotonic along the evidence sequence, with mid-trajectory failures that endpoint-only evaluation cannot detect. All scenarios, references, prompts, metrics, and analysis code are released for offline, hardware-free reproduction.
Read the full abstract