Peer Revenue Returns subject to peer review

Papers

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

All of arXiv

Monday 31 August

Computer Vision

AI-enabled Low-Cost 3D Maize Ear Morphometry Platform at Breeding Scale

Maize ear geometry (length, width, curvature, and volume) is closely tied to yield and grain-filling outcomes, but existing high-throughput phenotyping pipelines remain constrained by the cost, labor, and specialized hardware they require. We developed and validated a low-cost pipeline that reconstructs a watertight 3-D mesh of a maize ear from a single 20-second video captured with a consumer-grade DSLR on a motorized turntable under uniform LED illumination. Camera poses from a multi-seed COLMAP procedure initialize a Neural Radiance Field (NeRF), and a cylindrical holder of known diameter, visible in every frame, provides automatic metric scaling with downstream geometric quality control. Applied to 300 ears spanning a diverse maize inbred panel, 250 (83.3%) passed automated processing and quality control. Skeleton length agreed with manual caliper measurements across all 250 ears (R^2 = 0.964, RMSE = 4.68 mm), and convex-hull volume agreed with water-displacement volume on a 15-ear subset spanning the full size range (R^2 = 0.982, RMSE = 5.26 mL). Residual length error grew with ear curvature, whereas bounding-box height, which records the same straight-line chord as calipers, showed no such trend; the discrepancy therefore originates in the measurement definition, since calipers record the chord while skeleton length traces the geodesic arc. The capture hardware costs approximately 607 USD, and operator involvement fell from roughly five minutes to one minute per ear, with all downstream processing running unattended. The platform provides a foundation for breeding-scale 3-D ear phenotyping.
Read the full abstract

All authors Therin Young, Elijah Rodriguez, Lisa Coffey, Talukder Zaki Jubery, Adarsh Krishnamurthy, Patrick Schnable, Baskar Ganapathysubramanian

High Energy Physics (Theory)

From Self-Dual to Physical $\mathbb{C}P^{N-1}$: Anomalies, Boundary Stokes Phenomenon, and Global Structure of $\theta$-vacua

We introduce a two-coupling generalization of $\mathbb{C}P^{N-1}$ model that continuously interpolates between the self-dual ($\epsilon=0$) and the physical ($\epsilon=g$) theories as a useful nonperturbative tool. At $\epsilon \neq g$, this model possesses a chiral imbalance, which may be viewed as a real topological deformation (imaginary-$\theta$). We demonstrate that exact quantum equivalence between first- and second-order formulations strictly requires a topological counterterm sourced by a bosonic chiral anomaly. Solving this deformed theory at large $N$ yields two primary results. First, we analytically determine the nonperturbative vacuum structure of the self-dual theory, a self-dual vacuum with a dynamically generated field-strength condensate. Second, we resolve a fundamental paradox where saddles with $(\theta + 2\pi n) \sim O(N)$ ($n$ is branch number) spuriously yield lower energy densities than the physical ground state. Because the effective action possesses an essential singularity at $F=0$, we show that the Lefschetz thimble analysis must be generalized to include boundary thimbles. A boundary Stokes phenomenon renders the problematic saddles topologically inactive, fully restoring the validity of the large-$N$ expansion for strongly coupled theories.
Read the full abstract

Distributed Computing

Projection-Free Bandit Online Optimization for Multi-Agent Systems with Dynamic Regret

This paper investigates distributed online optimization for multi-agent dynamical systems with constrained inputs and time-varying cost functions. While online convex optimization offers a principal framework for sequential decision-making, existing online learning and optimization algorithms typically require accurate system models, limiting their applicability in practical settings. To overcome this challenge, we propose a distributed bandit online feedback optimization algorithm that relies solely on real-time input-output data. The algorithm employs a smoothing zeroth-order one-point estimator to construct local gradient approximations directly from cost evaluations. Additionally, to enforce input constraints effectively, we integrate a projection-free conditional gradient update, making the algorithm well-suited for online and large-scale settings. Furthermore, we establish a sublinear dynamic regret bound that depends on a temporal variation measure of system non-stationarity. Finally, numerical simulations demonstrate the effectiveness of the proposed algorithm.
Read the full abstract

Computation and Language

CPR for LLMs: Critical-Point Routing against Catastrophic Forgetting in Domain Adaptation

Supervised fine-tuning (SFT) is the de facto standard for adapting large language models (LLMs) to target domains, but it often degrades the model's general capabilities, a phenomenon known as catastrophic forgetting. Existing approaches typically modify the SFT loss to mitigate forgetting, but they inevitably operate along a domain-generality trade-off. In this work, we step outside this trade-off by decoupling the two capabilities at the model level: we keep the original base model for general capability, and selectively invoke the SFT expert only when domain-specific knowledge is required. Specifically, we propose CPR (Critical-Point Routing), a token-level routing framework between a base model and its expert derivative, based on critical tokens where the base model fails but the expert succeeds. We train a lightweight hierarchical router that estimates the expert-call probability per token, and pair it with a tailored inference procedure that combines momentum smoothing and threshold gating. Across diverse model-domain configurations, CPR achieves state-of-the-art across all settings, surpassing SFT expert by 1.4-5.5% in domain performance while recovering its general-capability drop from 3.4-14.5% to at most 0.5%, with minimal overhead from invoking the expert on only one-third of tokens.
Read the full abstract

All authors Kwangmin Ki, Yunhun Nam, Jongheon Jeong, Jaehyung Kim

Systems and Control

Multi-Domain Graph-Based Modeling of Energy Systems with Applications to Lithium-Ion Batteries

Graph-based models have been shown to provide a structured representation for complex multi-domain energy systems but face limitations when edge power flows depend on non-adjacent states or when a single edge carries multiple power-flow types driven by different inputs. This paper proposes two general extensions to address these limitations: a recursive state-to-input feedback scheme that embeds non-adjacent state dependencies into edge inputs without altering the graph structure, and a parallel edge decomposition method that represents composite interactions using sets of single-input edges while preserving energy conservation at the vertices. The extended framework is demonstrated on a lithium-ion battery module consisting of 36 parallel cells, and the resulting model predicts module temperatures with errors below 1{\deg}C. Validation on this electro-thermal battery system demonstrates the effectiveness of the extended framework for multi-domain systems that cannot be represented by previously established graph-based formulations, and indicates its potential for broader application to complex energy systems in control and design studies.
Read the full abstract

Computation and Language

Reactivating Test-Time Scaling for Plane Geometry Problem Solving

Plane geometry problem (PGP) solving has become a critical benchmark for multimodal reasoning because it requires accurate visual perception and precise multi-step symbolic deduction. Although test-time scaling (TTS) has demonstrated remarkable success in general mathematical reasoning, it fails to scale effectively under the symbolic-program paradigm for plane geometry. We identify two key obstacles: limited reasoning diversity induced by rigid symbolic programs and insufficient explicit visual grounding before symbolic deduction. To address these issues, we propose Multi-Trace Synthesis (MTS), which converts each symbolic program into heterogeneous reasoning traces, including executable Python scripts and CoT-augmented variants. We further propose Perception-Augmented (PA) training, which parses diagrams into structured semantic clauses before deduction, and Consensus-Guided Multi-Trace Ensemble (CG-MTE) for efficient self-adaptive inference. Experiments on three geometry benchmarks show that our method consistently improves PGP-solving across model scales and achieves strong performance against both general-purpose MLLMs and specialized geometry solvers. Under test-time scaling, CG-MTE achieves comparable accuracy to high-budget self-consistency while reducing sampling cost by up to 8x. Code and data are publicly available at https://github.com/Jason8Kang/ReTTS-PGPS.
Read the full abstract

All authors Xiaoqiang Kang, Shengen Wu, Maizhen Ning, Xiaobo Jin, Kaizhu Huang, Yutao Yue, Xiaowei Huang, Qiufeng Wang

Analysis of PDEs

Gradient Estimates Near the Natural Exponent for Very Weak Solutions to $A_p$-Weighted Quasilinear Elliptic Equations

We establish local Calderón–Zygmund estimates near the natural exponent for very weak solutions to matrix-weighted quasilinear equations\[ - \mathrm{div\,} A_{\mathbb{M}}(x,Du) = - \mathrm{div\,} A_{\mathbb{M}}(x,\mathbf{f}), \qquad A_{\mathbb{M}}(x,\xi)=\mathbb{M}(x)A(x,\mathbb{M}(x)\xi), \] where $A$ has $p$-growth and strong monotonicity, and $\mathbb{M}$ is a measurable positive-definite matrix field. We assume that $\mathbb{M}$ has bounded condition number and that \(\omega:=|\mathbb{M}|^p\in A_p\), without imposing uniform upper or lower bounds on $\mathbb{M}$. This extends the near-natural Calderón–Zygmund theory developed by Adimurthi–Phuc \cite{AP15} to the matrix-degenerate setting: There exists $\delta_0>0$ such that every very weak solution $u \in W^{p-\delta_{0}}_{\omega, \mathrm{loc}}$ satisfies \[ \mathbf{f}\in L^\gamma_{\omega,\mathrm{loc}} \Longrightarrow Du\in L^\gamma_{\omega,\mathrm{loc}} \] for $p-\delta_0\le\gamma\le p+\delta_0$. The proof requires handling the lack of energy estimates below the natural exponent and the use of Lipschitz truncation in the weighted setting. We achieve this through comparison estimates, higher integrability, and weighted analysis techniques.
Read the full abstract

General Relativity and Quantum Cosmology

Spacelike reduction of the gravitational topological terms and associated helicity densities

Helicity in gravity is a multifaceted concept, and several inequivalent definitions exist in the literature. In this work, we reduce the topological Pontryagin and Nieh-Yan terms to a constant time spacelike hypersurface. Remarkably, upon applying the SVT decomposition at the linearized level, we obtain four distinct helicity densities. The helicity density arising from the Nieh-Yan term is termed the spin-1 and spin-2 gravitomagnetic helicity density, depending on whether it originates from vector or tensor modes. Meanwhile, the helicity density arising from the Pontryagin term is termed the spin-1 and spin-2 gravito-current helicity density, according to the corresponding mode contributions. We study the mode and multipole expansions of these helicity functionals and apply them to leading order Newtonian two-body systems, weak field boosted Kerr, a locally defined slowly varying gyratonic pp-wave model, and a linearized gravitational Hopfion.
Read the full abstract

All authors Hongxi Huang, Jiang Long, Gan Zhao, Xin-Hao Zhou

Complex Variables

Transcendental Morse inequality on Kähler manifolds

We prove the transcendental Morse inequality conjecture of Boucksom-Demailly-Păun-Peternell. Among the corollaries of this result, this proves orthogonality of divisorial Zariski decompositions, shows that the pseudoeffective and movable cones are dual, and establishes the differentiability of the volume function on the big cone, on all compact Kähler manifolds.
Read the full abstract

Machine Learning

Converse and Collision-Based Achievability for Node Localization with Hybrid Distance-Spectral Graph Positional Encodings

Graph positional encodings are widely used in graph neural networks and graph Transformers, yet it remains unclear when the code itself can identify nodes. We study a hybrid distance-spectral encoding that combines anchor-distance profiles with quantized low-frequency Laplacian-energy coordinates. Treating the encoding as an observation map yields a simplex-refined converse, an exact collision factorization \(\kappa_H=\kappa_D\kappa_{S|D}\), and the collision information \(I_H=-\log\kappa_D-\log\kappa_{S|D}\). On random regular graphs, the criterion is made explicit through a bounded-correlation Gaussian-wave surrogate; for actual Laplacian-energy coordinates, we give the distance-conditioned spectral collision condition sufficient for conditional actual-coordinate achievability. Experiments show that \(I_H/\log n\) calibrates localization success, and PE-only structural task probes on Universal Dependencies trees show that hybrid encodings better recover syntactic-tree geometry than distance-only or spectral-only baselines.
Read the full abstract

All authors Zimo Yan, Yifan Li, Hao Li, Zheng Xie, Chang Liu, Zheming Tu, Yuan Wang

Soft Condensed Matter

Linear stability and viscoelastic response in weakly jammed frictional granular systems

Granular particles typically feature contact friction, which gives rise to various nontrivial properties of granular materials. In this work, we investigate the linear stability of static packings composed of frictional spheres, whose interactions are modeled using the discrete element method. We discuss a key difficulty in properly modeling static friction and address the problem of identifying rattlers. We also explore the spectrum of the energy matrix, which governs the linear viscoelastic response of the system. The intermediate- and high-frequency regimes of the spectrum remain qualitatively independent of both the friction coefficient and the packing fraction.
Read the full abstract

All authors Masanari Shimada, Yusuke Hara, Shihori Koyama, Ryohei Seto, Maria Yokota

Mesoscale and Nanoscale Physics

Universal tuning of Förster resonance energy transfer in gate-programmable conductor-dielectric-conductor heterostructures

We develop a quantum-electrodynamical theory for universal tuning of spontaneous emission and Förster resonance energy transfer (FRET) in a material-agnostic conductor-dielectric-conductor heterostructure. The platform consists of a dielectric spacer of thickness $W$ bounded by two gate-tunable two-dimensional conductors. The only microscopic input from the surrounding materials is the transverse-magnetic and transverse-electric reflection amplitudes $r_{\TM/\TE}(q_\rho,\omega)$ of the two sheets. Starting from the QED photon propagator, we derive the retarded Maxwell dyadic, the vacuum/Hadamard field propagator, and the time-ordered Feynman propagator in the same geometry. The spontaneous-emission rate is controlled by the local vacuum spectral density, while FRET is controlled by the nonlocal retarded/advanced product $\obG_{\text R}(\bx_D,\bx_A;\omega_D)\Im\,\balpha_A(\omega_D)\obG_{\text A}(\bx_A,\bx_D;\omega_D)$. In the transparent limit $r_{\TM}\to 0$, the near-field FRET rate recovers the bulk $x_\rho^{-6}$ law. In the Dirichlet/PEC branch $r_{\TM}\to -1$, the gapless transverse mode is removed and the donor-acceptor coupling acquires a Bessel-$K$ envelope, giving an exponentially screened FRET rate $\Gamma_{D\to A}\propto\exp(-2\pi x_\rho/W)$ at large lateral separation. In the opposite Neumann/PMC-like branch $r_{\TM}\to 1^{-}$, a nearly gapless transverse mode survives and produces a wide quasi-two-dimensional logarithmic propagator, enhancing the nonlocal electromagnetic coupling over a gate-programmable range $x_{\rho,*}\sim W/(1-r_{\TM})$. For graphene-Er implementations, the same retarded Green tensor also separates dissipative on-shell Er-to-graphene decay, governed by its absorptive part, from dispersive virtual-plasmon-mediated Er-Er coupling, governed by its reactive part; a plasmonic band gap can suppress the former while retaining the latter.
Read the full abstract

Computer Vision

CedarCypress3D: an annotated UAV-LiDAR dataset of individual trees in planted cedar and cypress forests

Individual tree measurements derived from Light Detection and Ranging (LiDAR) mounted on Unmanned Aerial Vehicles (UAV) provide valuable information for forest inventory, ecosystem monitoring, and sustainable forest management. Recent advancements in machine learning have increased the demand for annotated datasets to develop and evaluate point cloud-based approaches, especially for individual tree segmentation. However, publicly available annotated UAV-LiDAR datasets in temperate forests are limited. In this article, we present CedarCypress3D, a manually annotated UAV-LiDAR dataset collected in Japanese cedar (Cryptomeria japonica) and Japanese cypress (Chamaecyparis obtusa) plantations in Japan. The dataset consists of UAV-LiDAR point clouds and field survey measurements from 34 circular plots across two sites with different topographic characteristics, along with terrestrial LiDAR point clouds available for a subset of 22 plots. A total of 1,627 trees were measured in the census field survey and manually annotated to match the corresponding trees in the UAV-LiDAR point clouds. For the subset of plots with terrestrial LiDAR data, semantic labels (i.e., stem and non-stem) were additionally assigned to tree points in the UAV-LiDAR data. CedarCypress3D provides high-quality annotated UAV-LiDAR data for developing and evaluating individual tree instance segmentation and semantic segmentation methods in temperate planted forests. The dataset can also support research on tree attribute prediction and multi-platform LiDAR analysis. The dataset is publicly available at https://doi.org/10.5281/zenodo.22168721.
Read the full abstract

All authors Katsuto Shimizu, Fumiaki Kitahara, Tomohiro Nishizono, Hideki Saito, Masayoshi Takahashi, Shingo Obata, Shunsuke Tei, Naoyuki Furuya, Tomoya Goto, Eiji Kodani, Yusuke Yamada

Algebraic Geometry

Tropical Matsusaka and Matsusaka–Ran Criteria

The tropical Schottky problem asks which principally polarized tropical abelian varieties are isomorphic to tropical Jacobian varieties equipped with their canonical principal polarizations. We prove the tropical Matsusaka criterion and the tropical Matsusaka–Ran criterion, which are the main results of this paper, and show that their hypotheses are equivalent. Together with the tropical Poincaré formula proved by Gross–Shokrieh, these results give a geometric solution to the tropical Schottky problem. In particular, we give an affirmative answer to a question posed by Brannetti–Melo–Viviani. Several tools are developed for the proofs of the two criteria. In particular, we establish Hodge-type inequalities for ample tropical Cartier divisors on tropical abelian varieties and prove a splitting criterion for principally polarized tropical abelian varieties.
Read the full abstract

Computation and Language

CAST: Critique-Aware Supervision for Training Reliable Long-Horizon Tool-Calling Agents

Large language model (LLM) agents are increasingly deployed in long-horizon, interactive, and stateful environments. In these settings, a single wrong action, such as refunding the wrong purchase, can cause irreversible task failure and must be intercepted before execution. Such failures may not appear in every single run, but can emerge across repeated trials, making reliability across steps and trials critical. However, ensuring agentic reliability is challenging: even frontier LLMs struggle to explain why an action may be wrong, especially in long, intertwined trajectories governed by domain-specific policies. Much recent work relies on prompt-based critique agents, while optimization-based methods lack a systematic way to produce rich verification rationales for training. We address this gap with CAST, a critique-aware training framework that converts sparse task outcomes into action-level supervision for critique learning and policy optimization. CAST analyzes agent trajectories to synthesize structured rationales explaining action validity under partial observability. The resulting critique model is used to construct critique-aware training data for optimizing the policy model. Fine-tuning Qwen3-family models on dynamic tool-calling benchmarks, CAST improves reliability across domains, outperforming GPT-OSS-120B by over 10% pass^4 on Retail tasks and yielding an additional 9% improvement on Telehealth in an out-of-domain setting. These results demonstrate that critique-aware training improves the robustness of LLM agents in realistic dynamic environments.
Read the full abstract

All authors Amir Saeidi, Zehua Zhang, Rishitosh Singh, Naman Ahuja, Vivek Gupta, Ali Payani, Gaowen Liu, Jayanth Srinivasa, Chitta Baral

Optics

Revealing antenna-molecule coupling using virtual gain from synthetic complex frequency analysis

Synthetic complex frequency waves (CFWs) enhance spectral and spatial resolution by compensating loss through virtual gain. This approach is particularly useful in nano-photonics where loss is intrinsically linked to strong field confinement and system size. Here, we apply CFWs to metasurfaces composed of Fano-resonant plasmonic antennas with lithographically tunable coupling between the weak and strong coupling regimes and where loss may be reduced to the Drude damping limit. We quantitatively assess the impact of the virtual gain to find that, while reducing loss, internal coupling rates remain invariant. While the exceptional point associated with strong light-matter coupling is invariant under virtual gain, the cooperativity of Fano interference is enhanced. When the metasurfaces are functionalised with a molecular monolayer, CFWs are thus able to reveal a molecule's vibrational modes, even when obscured by loss, and allow direct extraction of the plasmon-molecule coupling strength. We thus use CFW analysis as a quantitative tool for revealing and assessing light matter-coupling.
Read the full abstract

All authors Toby Severs Millard, Xiaofei Xiao, Mayela Romero-Gómez, Tom Aktaş-Vincent, Nathaniel J. Huáng, Chris C. Phillips, Stefan A. Maier, Olga Kazakova, Rupert F. Oulton

Information Retrieval

Understanding before verifying: Claim normalization for automated citation verification

Citation accuracy has been studied for decades because of its importance to research reliability. Content-level citation verification assesses the reliability of scholarly claims. Recent work adopts a two-stage retrieval-classification framework inherited from fact-checking. However, this design overlooks the complexity of the raw citing claim and introduces three issues into the verification system, namely scope mismatch, perspective mismatch, and proposition entanglement. These issues increase the difficulty of retrieval and classification, thereby limiting model performance. Motivated by this gap, we propose claim normalization, which applies three rewriting strategies to the raw citing claim before retrieval and classification, allowing each downstream model to perform a single, well-defined task. Building on this method, we develop Claim-Normalized Citation Verification (CNCV), a new three-stage framework consisting of claim normalization, evidence retrieval with grounding, and citation classification. We evaluate CNCV across 18 classifiers using a factorial experiment on human-annotated citation instances. Compared with the prior two-stage framework, CNCV improves macro F1 by an average of 12% for encoders and 10% for generative LLMs, driven by improved evidence quality, the dominant factor identified in our experiments. Evidence retrieved from automatically normalized claims yields downstream classification performance statistically equivalent to that obtained with manually annotated evidence.
Read the full abstract

All authors Yifan He, Mengjia Wu, Siming Deng, Yi Zhang

Robotics

Rethinking Language's Role in Efficient VLA for Autonomous Vehicles: Toward Smarter, Trustworthy Driving

Vision-Language-Action (VLA) models are reshaping autonomous driving (AD) by unifying perception, reasoning, and control through language, enabling semantic grounding, interpretable decisions, and better long-tail generalization. But language is expensive onboard: latency and memory budgets are tight, and autoregressive decoding is inherently sequential. This work reframes the central question as when and where language should act at inference, since inference cost recurs at every deployed frame while training cost is paid once. We introduce the Language Residue taxonomy to organize methods by their inference-time use of language: train-time-only supervision (L1), latent non-textual reasoning (L2), conditional invocation (L3), and full per-frame generation (L4). We review representative methods and tag each across five deployment axes (latency, parameters, memory, FLOPs, tokens), analyzing them on major open- and closed-loop driving benchmarks (e.g., nuScenes, NAVSIM, Bench2Drive). We further trace how efficient methods from NLP/LLM are adapted in AD, identifying the constraints and motivations driving these adaptations. A continuously updated repository will be available at Github.
Read the full abstract

Computer Vision

Repeatability Characterisation and Error Budget of a Consumer Structured-Light Scanner for 3-D Wound Geometry:A Rigid-Phantom Study

We characterise the measurement error of a free- hand consumer structured-light scanner used to derive three- dimensional geometric wound descriptors. Rigid wound-care phantoms cannot change, so every difference between repeat scans of one site is measurement error; all repeats come from a single scanner unit, nine sites and 23 scans, so this charac- terises one instrument. The 95 percent repeatability limit for reconstructed surface area is a factor of 4.8, with a confidence interval from 3.0 to 7.0, so rescanning an unchanged site can shift the reading from a 79 percent decrease to a 377 percent increase. Forty-four of 45 descriptors fall below an intraclass correlation of 0.50, and none of 248 descriptor and pipeline combinations reaches 0.75. An error budget formed by holding the analy- sis region fixed, leaving sensor and reconstruction untouched, bounds the share of variance attributable to how much surface the operator captured at 75 percent for surface area, 91 percent for hull area and 95 percent for bounding-box diagonal; only hull volume is majority instrumental, at 48 percent, so a better sensor would buy little. No wound is delineated anywhere in the chain, so the comparison against the four-week area reduction used clinically to predict healing, a ratio near 2.1, is a lower bound on the noise an unsegmented pipeline must overcome, not a measurement of wound-area reproducibility. Standardising the analysis region cuts the limit to 2.13, meeting that ratio rather than clearing it. Statistical outlier removal imposes a measured systematic area deficit near 11 percent.
Read the full abstract

Human-Computer Interaction

LandmarkLens: Predicting and Presenting Effective Landmarks for Mixed-Reality Urban Exploration

People with a poor sense of direction (SOD) struggle to build cognitive maps for effective spatial navigation, and existing navigation tools prioritize efficiency over spatial learning. To understand how navigation strategies differ by ability, we conducted a landmark attention study with 20 participants (ten good SOD, ten poor SOD) who navigated across four Tokyo neighborhoods in virtual reality (VR). We found systematic group differences in both gaze behavior and the types of landmarks they verbally identify as effective. Based on these findings, we built LandmarkLens, a mixed-reality (MR) navigation system that uses a vision-language model (VLM) to identify and highlight navigation-relevant landmarks. A follow-up study with eight poor-SOD participants showed improved performance in scene recognition, suggesting that guided landmark attention can support landmark-level spatial knowledge acquisition for people with poor SOD, a first step toward broader spatial learning.
Read the full abstract

All authors Chu Li, Yotam Sechayk, Jared Hwang, Jon E. Froehlich, Takeo Igarashi

Cryptography and Security

Balancing Privacy, Utility, and Safety in LLM Alignment through Preference Optimization

Preference optimization is widely used to align large language models with human preferences, but preference-data composition may also influence privacy-relevant memorization. We examine whether adding synthetic privacy-preference pairs to Direct Preference Optimization (DPO) is associated with lower canary-based memorization signals without modifying the objective or introducing a formal privacy mechanism. We propose Privacy-Pressure Preference Mixing (P3M), a data-composition protocol that varies the amount of privacy-preference data while keeping helpfulness and harmlessness preference data fixed. We evaluate a non-privacy Baseline and privacy-mixing ratios of 0.5, 1.0, and 2.0 using Gemma 3 270M-IT across five random seeds and validate the same four conditions using 4-bit-quantized Gemma 2 2B-IT across three seeds. Overall, under the tested conditions, privacy-preference mixing is associated with lower mean canary suffix log-likelihood proxy values across both model settings and lower aggregate membership-inference attack performance relative to the Baseline in the mixed-source 2B evaluation. Specifically, across the privacy-aware 2B configurations, the mean area under the receiver operating characteristic curve (AUROC) ranges from 0.596 to 0.629, and the mean area under the precision-recall curve (AUPRC) ranges from 0.541 to 0.575, compared with 0.804 and 0.790, respectively, for the Baseline. However, the reduction in membership distinguishability does not hold uniformly across data sources. Moreover, the relationship between the privacy ratio and harmlessness preference accuracy varies by model setting, whereas helpfulness preference accuracy remains broadly stable. These findings suggest that P3M should be viewed as a lightweight empirical protocol for examining privacy-utility-safety trade-offs rather than as a formal privacy guarantee or a defense against extraction attacks.
Read the full abstract

Earth and Planetary Astrophysics

Empirical Constraints on the CO Snowline Transition in HD 163296: The Local CO Column and $^{13}$C$^{18}$O Optical Depth

CO isotopologue emission is widely used to infer gas masses and volatile carbon abundances in protoplanetary disks, but converting line emission into a CO column depends on optical depth, temperature structure, linewidth, and isotope ratios. Inferring CO/H$_2$ additionally requires an independent constraint on the local hydrogen column. We use high-resolution $^{13}$C$^{18}$O $2$-$1$ observations of HD 163296 to derive a spatially localized empirical constraint on the CO column at the resolved CO snowline edge. We focus on the 70-75 au annulus, on the inner, high-column side of the observed profile steepening near 75 au. Using RADEX slab calculations conditioned on a two-dimensional temperature structure, we infer an effective beam-averaged CO column from the absolute integrated intensity. Across representative temperatures, isotope-ratio pairs, and effective local linewidths of 0.30 and 0.50 km s$^{-1}$, we find $N_{\rm CO}^{\rm beam}=(1.6-2.4)\times10^{20}$ cm$^{-2}$ and $^{13}$C$^{18}$O line-center optical depths $\tau=0.39-0.97$. Thus, even this rare isotopologue is not safely optically thin at the snowline edge. Adopting $N_{\rm H2}=2.5\times10^{24}$ cm$^{-2}$ from a published parametric gas surface-density profile gives the conditional abundance ${\rm CO/H_2}=(6.4-9.5)\times10^{-5}$. A beam-forward radial-profile analysis gives consistent columns, and a physical disk model with a near-canonical warm-layer CO abundance supplies a comparable CO column. The measurement is consistent with the higher C$^{17}$O-based MAPS estimate. For the adopted hydrogen column, the inferred CO/H$_2$ ratio is consistent with a near-canonical abundance on the warm side of the snowline.
Read the full abstract

All authors Chunhua Qi, Takahiro Ueda, David J. Wilner, Catherine C. Espaillat

Superconductivity

Superconductivity of Tellurium Polyhydride with Tc above 90K

We report experimental diacovery of superconductivity (SC) in tellurium (Te) polyhydride. The compound was synthesized at high pressure and high temperature conditions using a diamond anvil cell combined with a laser heating system. Subsequent in situ transport measurements at high pressures, performed as a function of temperature and applied magnetic field, revealed a superconducting transition with a critical temperature Tc about 91 K at 263 GPa. The superconducting phase is assigned to TeH4 with characterized face shared TeH12 cage forming quasi molecular H2 units based on synchrotron x-ray diffraction experiments. Analysis of the SC behavior at magnetic fields yielded a Ginzburg Landau (GL) coherence length of approximately 54 angstroms. Tellurium polyhydride thus becomes another chalcogen polyhydride superconductor in addition to the landmark discovery of the first polyhydride high Tc SC SH3.
Read the full abstract

All authors Jinfu Zhu, Guiqi Liu, Yuanhao Su, Hongyu Liu, Sijia Zhang, Panpan Kong, Qingqing Liu, Jianfa Zhao, Shaomin Feng, Jun Zhang, Haoyu Zheng, Jing Song, Luhong Wang, Fuyang Liu, Haozhe Liu, M. Bykov, Xiancheng Wang, Changqing Jin

Analysis of PDEs

High frequency wave propagation for the viscoelastic wave equation with singular memory

We study high-frequency propagation for a viscoelastic wave equation with spatially dependent hereditary memory in relative-history form. The kernel may have the integrable singularity $\mathfrak m(s,x)=s^{p-1}m(s,x)$, $0<p<1$; the regular case $p=1$ is included. Using the half-amplitude propagation distance and corresponding travel time as units, the wavelength $h\ll1$ yields the memory factor $\varepsilon=h^{1-p}$. We construct exact solutions with full two-scale geometric-optics expansions in powers $h^{k+(1-p)\ell}$. Memory modifies the propagation geometry through the instantaneous modulus $\sigma+\varepsilon\int_0^\infty\mathfrak m(s,\cdot)\,d s$, while the kernel singularity contributes $C_p=\Gamma(p)e^{i\pi p/2}$ to the leading transport equation. For $0<p<1$, this produces fractional scales, frequency-dependent attenuation, and a dispersive phase correction; for $p=1$, the fractional hierarchy disappears, attenuation is frequency independent, and the transport phase correction vanishes. We also derive a local damped wave equation whose incoming high-frequency solutions approximate the hereditary solutions with $O(h)$ error in semiclassical $C^k$ norms. Exterior observations for all incident directions and $0<h\ll1$ uniquely recover $\sigma_{\mathfrak m}$ and the full temporal jet of $m$ at $s=0$, which determine the expansion modulo $O(h^\infty)$. Finally, a contraction-semigroup argument gives well-posedness and arbitrary finite-order Sobolev regularity for spatially dependent weakly singular kernels and prescribed full prehistory, with explicit compatibility conditions and estimates uniform in $\varepsilon$. These estimates justify the geometric-optics construction.
Read the full abstract

Optimization and Control

Enhancing Interpretability of Stochastic Programming Solutions: A Multiparametric Approach

Stochastic programming (SP) is a powerful framework for decision-making under uncertainty, but its practical adoption in industry is often hindered by the difficulty in understanding the causal relationships that drive optimal solutions. In the two-stage SP, strategic first-stage decisions are coupled with operational second-stage recourse decisions. When the number of scenarios under consideration is large, understanding the direct link between the uncertainty realization and optimal recourse strategy becomes computationally and cognitively demanding. Common approaches to improve interpretability include trained classification trees or scenario reduction, replacing the large scenario set with a representative subset. This is often achieved through post-hoc clustering (e.g., k-means) based on uncertainty realizations or optimal recourse decisions. While useful, these methods only provide a statistical approximation of the solution space and may fail to reveal the underlying structural properties of the recourse problem that drive optimal first-stage decisions. This work introduces a novel, deterministic approach to explainability using multiparametric programming (mp) within a Benders decomposition framework. We reformulate the recourse subproblem as a multiparametric linear program, generating an explicit map of Critical Regions (CRs), which are polyhedral partitions of the uncertainty space. This allows us to cluster scenarios analytically rather than statistically. We demonstrate this methodology on a supply chain planning under demand uncertainty. Our results show that 100 stochastic scenarios map to exactly six critical region clusters. This mapping allows us to explain optimal capacity planning decisions as a precise trade-off between specific operational modes, providing a fully transparent interpretation of the stochastic solution.
Read the full abstract

Optimization and Control

Multi-Agent Receding Horizon Games Framework for Autonomous Market Participation

Electrochemical distributed energy resources (DERs), such as electrolyzers and battery energy storage systems, consume large amounts of electricity. Buying that power directly from wholesale markets can sharply reduce operating costs. Accessing these markets, however, requires a minimum bid size that individual small or medium-scale units struggle to meet. Grouping several units to bid as a single participant clears this barrier while letting each unit use its own low-cost renewable generation. The common remedy is to hire a third-party aggregator, but aggregators charge commissions and solve a centralized optimization that often favors certain units over others. A fairer alternative is peer-to-peer (P2P) participation, where units form a self-governing group and bid jointly with no central authority. Because each self-interested unit is unwilling to share private data, coordination is best posed as a game-theoretic distributed optimization problem. Existing P2P methods, however, address only single-round spot markets and ignore the two-stage structure of real wholesale markets, where participants commit a day ahead and continuously adjust in real time. We close this gap with a two-stage receding-horizon generalized Nash equilibrium (GNE) game. Each unit re-optimizes its strategy every five minutes over a rolling one-hour horizon, while the group satisfies both market stages collectively. Units exchange only publicly visible aggregate power, never private cost or production data. We apply the approach to a six-unit fleet on the PJM market. It delivers 37% higher profit, 29% lower electricity cost, and 92% less renewable curtailment than individual participation, with every unit better off.
Read the full abstract

Computation and Language

Verification-Aware Training for Speculative Decoding

Speculative decoding accelerates large language model inference by using a draft model to generate candidate tokens, which are verified by the target model in a single forward pass. Verification proceeds sequentially and discards every position from the first rejection onward, yet existing draft training relies on token-level imitation of the target with a fixed per-position weighting that reflects neither property. We introduce Verification-Aware Training (VAT), a plug-in framework that simulates verification at every training step and turns the resulting accept and reject patterns into supervision. VAT consists of two components: (i) a verification head, a lightweight jointly trained binary classifier that supervises the draft model on whether each position survives sequential verification; (ii) verification-adaptive weighting, which replaces the fixed weighting schedule by keeping full weight up to each sample's first rejection point and re-anchoring the decay to start there. VAT modifies only the training objective, so it can be layered on top of existing methods without changing the draft architecture, the target model, or the inference procedure. Applied to EAGLE-3 and DFlash on Qwen3-4B, Qwen3-8B, and LLaMA-3.1-8B, VAT improves average acceptance length by up to 11.4% and wall-clock speedup by up to 8.7%, with consistent gains across math, code, and chat benchmarks. Code will be available at https://github.com/naver-ai/vat
Read the full abstract

All authors Geonmo Gu, Byeongho Heo, HeeJae Jun, Yoohoon Kang, Sangmin Lee, Sangdoo Yun, Dongyoon Han

Optimization and Control

Solution and Optimal Strategies for a Differential Game of Pursuit-Evasion with Point-Wise Constraints

This work study a differential game of pursuit-evasion involving countably many pursuers $p_{1},p_{2},\cdots,p_{m}$ and a single evader $e$ under point-wise (geometric) constraints in the sequence space $l_{2}.$ The pursuers evolve according to specified differential equations of $1^{st}$ order while the evader follows a $2^{nd}$ order differential equation. The game period is a fixed time interval of length $\theta$ unit of time. The minimum distance between evader and the pursuer at the terminal time denotes the game payoff. The pursuers aim is to minimize the payoff, whereas the evader seeks to maximize it. We obtain the solution of the game, including optimal strategies of the players construction, as well as game value estimation.
Read the full abstract

All authors Jamilu Adamu, Abbas Ja'afaru Badakaya, Felix Wallace Tarry, Mehdi Salimi

Materials Science

Attosecond Reconstruction of Strain Tensors via Electronic Fingerprints

We demonstrate an attosecond transient absorption spectroscopy (ATAS) scheme for reconstructing strain tensors in two-dimensional materials. Using strained graphene as a prototype system, we show that the fishbone structures in ATAS serve as distinctive spectral fingerprints of strain, where strain-induced shifts and splittings of van Hove singularities encode the magnitude and orientation of the strain tensor, respectively. By combining density-matrix simulations with analytical modeling, we establish a direct mapping between transient absorption spectra and strain tensors, enabling accurate retrieval of lattice deformation from ultrafast electronic responses. Our work introduces an attosecond spectroscopic paradigm for ultrafast strain metrology, where electronic fingerprints replace conventional structural probes for sensing lattice deformation in quantum materials.
Read the full abstract

All authors Jing Li, Jiayu Yan, Guoyong Yuan, Chao Chen, Shang Wang, Fulong Dong

Geometric Topology

R-equivalence of quandle colorings and inner automorphisms

R-equivalence is an equivalence relation on the set of colorings of an oriented knot diagram by a quandle. In this paper, we show that a certain subgroup of the inner automorphism group of a quandle acts on the R-equivalence class of a given coloring by the quandle. We also determine the R-equivalence classes of colorings of a diagram of a $2$-bridge knot by a dihedral quandle completely, under a certain condition.
Read the full abstract