Search arXiv⌕ Search

arXiv subjects

Yuan Gao

Publications and source records attributed to Yuan Gao.

At least 19 recordsLinked to original sources
arXiv2026

Hidden Triakontadipole Order in the Triangular-Lattice Magnet CeMgAl$_{11}$O$_{19}$

The absence of magnetic Bragg peaks and a broad excitation continuum are commonly recognized experimental features of a quantum spin liquid. The Ce-based triangular-lattice hexaaluminate CeMgAl$_{11}$O$_{19}$ exhibits both signatures, yet its effective model and ground state remain under debate. Experiments have variously suggested a quantum spin liquid, proximate umbrella point, and disorder-suppressed magnetic order. Here we trace these contradictions to an overlooked multipolar degree of freedom: the transverse pseudospin of the $|\pm5/2\rangle$ Kramers doublet is a rank-5 triakontadipole rather than an ordinary dipole. Analyzing the magnetization, neutron spectra, and thermodynamics on equal footing, we establish a dipole-triakontadipole XXZ model with $J_{\perp} \simeq 0.65$ K and $J_c \simeq -0.13$ K. The resulting planar triakontadipolar superfluid order is hidden from dipolar probes, explaining the absence of a magnetic Bragg peak in neutron scattering. The zero-field excitations separate into a low-energy triakontadipole wave and a broad higher-energy continuum. We further propose nuclear spin-lattice relaxation and spin-current experiments to probe the gapless excitations of this hidden triakontadipole superfluid. Our findings thus open a new avenue for high-rank multipolar physics in rare-earth triangular-lattice magnets.

cond-mat.str-el↗
arXiv2026

Selection principle for Weak KAM solutions on $\mathbb{R}^n$ via large deviation principle for diffusions

Stationary Hamilton-Jacobi equations generally admit many weak KAM solutions. The question is how to select one weak KAM solution that reflects a meaningful physical principle. In this paper, we propose a selection principle for Hamilton-Jacobi equations arising from the large deviation principle of general drift-diffusion processes on $\mathbb{R}^n$ based on the exponential asymptotics of invariant measures. We first characterize the Aubry set as the finitely many equilibria of the drift, using an action estimate from the orthogonal decomposition of the drift. For any boundary data on the Aubry set compatible with the Peierls barrier, we construct the unique coercive weak KAM solution with these boundary data, which is also the maximal locally Lipschitz viscosity solution. The Freidlin-Wentzell graph formula then provides the boundary data on the Aubry set that select the unique weak KAM solution. The selected solution is exactly the rate function in the large deviation principle for invariant measures in the zero noise limit. We also show that this rate function is the vanishing viscosity limit, to which the viscous approximations converge locally uniformly. Finally, under additional regularity assumptions, we show that the long-time and zero-noise limits commute for suitably prepared initial densities.

math.AP↗
arXiv2026

Nearby Lagrangians in the cotangent bundle of the Klein bottle

We prove the nearby Lagrangian conjecture for the Klein bottle: every closed connected embedded exact Lagrangian submanifold of its cotangent bundle is Hamiltonian isotopic to the zero section by a compactly supported isotopy. The proof passes to the torus double cover and builds on the theorem of Dimitroglou Rizell, Goodman, and Ivrii for the two-dimensional torus. The main step is to construct an isotopy compatible with the covering involution. We achieve this using holomorphic sphere rulings in $S^2\times S^2$ adapted to the zero section, allowing the construction to descend to the Klein bottle.

math.SG↗
arXiv2026

Koszul duality for boundary categories

We study how Koszul duality relates two kinds of boundary constructions associated with a smooth category and a proper category. For a smooth category with a proper Koszul-dualizing subcategory, we construct a fully faithful functor from its categorical neighborhood of infinity to the singularity category associated to the proper category. We prove that this functor is an equivalence precisely when the smooth category is reflexive, meaning that it can be reconstructed from its finite-dimensional module theory. Under a Gorenstein condition, satisfied in the Calabi--Yau setting, the functor is induced by restriction of modules. We give algebraic criteria that ensure equivalence and obtain an identification of Hochschild cohomology.

math.KT↗
arXiv2026

On the Faithfulness of Visual Thinking: Measurement and Enhancement

Recent large vision-language models (LVLMs) can generate vision-text multimodal chain-of-thought (MCoT) traces after reinforcement fine-tuning (RFT). However, we observe that the visual information in MCoT is often inaccurate even when answers are correct, and task accuracy remains nearly unchanged despite the visual information being corrupted. This indicates a lack of faithfulness in the vision part of MCoT reasoning. We attribute this to the RL reward design in RFT, which solely incentivizes the format of interleaved vision-text cues, encouraging the model to incorporate visual information into its text reasoning steps without considering its correctness. In this paper, we first probe MCoT faithfulness by measuring how much the prediction changes when its visual and textual thoughts are intervened. Surprisingly, the model's predictions remain nearly unchanged under visual intervention but change significantly under textual intervention, indicating that the visual evidence is largely ignored. To further analyze the visual information, we introduce an automated LVLM-based evaluation metric that quantifies the faithfulness of visual cues from two perspectives: relevance and sufficiency. Our evaluation reveals that the visual information in current MCoT traces is simultaneously irrelevant and insufficient. To address this issue, we propose Sufficient-Component Cause Model (SCCM) learning. This approach encourages the MCoT to generate sufficient yet minimal visual components that are independently capable of leading to the correct answer. The proposed SCCM is annotation-free and plug-and-play, compatible with various RFT for MCoT. Empirical results demonstrate that SCCM substantially improves the visual faithfulness across a suite of fine-grained perception and reasoning benchmarks. Code is available at https://github.com/EugeneLiu01/Faithful_Thinking_with_Image.

cs.CV↗
arXiv2026

Edge intelligence for distributed acoustic sensing with ultracompact neural networks

Distributed acoustic sensing (DAS) transforms optical fibres into dense arrays of vibration sensors for continuous, long-range monitoring with high spatial resolution. Yet converting these large-scale measurements into actionable information with compact models that retain task performance remains challenging. Existing artificial-intelligence approaches are typically optimized for specific sensing environments and tasks, but their computational requirements can hinder deployment in compact, power-constrained field systems. Here, we introduce DAS-NanoNet, a framework that automatically constructs lightweight neural networks for different DAS scenarios and tasks, together with OmniDAS, an open benchmark spanning six representative environments. Using lightweight temporal operators tailored to DAS signals, an adaptive neural architecture search strategy derives a compact network for each scenario-task pairing. Across classification, localization, compression, and denoising tasks in diverse DAS applications, DAS-NanoNet consistently yields highly compact architectures. On a six-class dataset, it achieves 99.68% accuracy with only 1,391 trainable parameters, while delivering comparable or better performance with up to 18,098-fold fewer parameters than previously reported high-accuracy networks. Its compact architecture further enables edge deployment, achieving sub 9 $μ$s inference latency on the deployed hardware platform across different DAS applications. DAS-NanoNet provides a route towards edge intelligence for compact, low-power and real-time DAS systems through ultracompact networks tailored to individual sensing scenarios and tasks.

physics.optics↗
arXiv2026

Sticky Jump Diffusions: A Unifying View of Masked, Continuous, and Hybrid Diffusion

We introduce sticky jump diffusions (SJDs), continuous-time Markov processes on $\mathbb{R}^d$ whose discrete anchors are token embeddings. In forward time, anchors release their mass at a hazard rate and the released mass diffuses in the continuous ambient space; time reversal couples a score-driven SDE with a sticky jump kernel whose rate and destination are fixed by flux balance with the forward law. We estimate the score and the per-anchor reverse hazards from a single denoising classifier via denoising hazard matching, the hazard analogue of denoising score matching, with simulation-free cross-entropy training. SJD recovers masked diffusion, continuous diffusion, and hybrid diffusion as limits. Its reversal explains features that each family treats as given: the mask of masked diffusion carries no evidence about the source token because the unsticking kernel of every anchor collapses to the same absorbing point; the terminal projection of continuous diffusion is required due to the absence of atoms in its forward marginal, without which flux balance yields no reverse jumps; and the update rules of hybrid diffusion (commit rate, destination, and drift) all follow from flux balance rather than from separate design. Beyond these limits, the unsticking kernel becomes a design space: a cross-position blending corrupts each position toward a blend of its neighbors' clean values or embeddings, turning dependency structure such as spatial locality or a constraint graph into an inductive bias of the corruption itself, and improves over the identity-kernel hybrid on CIFAR-10, Text8, and Sudoku.

cs.LG↗
arXiv2026

From Solver Trajectories to Teaching Trajectories: Cognition-Aligned Reasoning Distillation

Knowledge distillation from large reasoning models (LRMs) often uses raw teacher trajectories as supervision. However, this practice creates a mismatch between the goal of distillation and the nature of teacher trajectories. Advanced LRMs typically improve their reasoning capabilities through reward-driven post-training, such as RLHF. This training optimizes them as strong solvers, rather than as teachers that produce reasoning processes smaller models can easily learn. Since smaller models have different "cognitive capacities" compared to their larger teachers, directly imitating raw teacher trajectories can sometimes be ineffective and often requires a substantial number of high-quality samples. To bridge this mismatch, we propose Cognition-Aligned Reasoning Distillation (CARD), an RL-based framework that transforms solver trajectories into teaching trajectories aligned with the student's cognition. Avoiding expensive resampling, our framework directly transforms existing trajectories. We optimize CARD with two reward signals: (1) a Distribution Alignment Reward that matches the student's predictive distribution, and (2) a Logical Coherence Reward that reduces misleading failed reasoning steps. Comprehensive evaluations across multiple student models and settings demonstrate that CARD outperforms distillation from raw teacher trajectories in in-domain settings while achieving substantial gains in out-of-distribution performance. Code is available at https://github.com/LiteSSLHub/CARD.

cs.AI↗
arXiv2026

KL-MOPD: Discrepancy-Aware Multi-Objective On-Policy Distillation for Generative Recommendation

Coordinating multiple recommendation objectives (e.g., click and conversion) is critical for real-world generative recommendation (GR). However, since a positive item in GR can be retrieved only when its tokens are successively selected at every decoding step, traditional item-wise methods lack fine-grained token-wise supervision to guide intermediate steps. Motivated by its success in foundation models, we consider introducing multi-objective on-policy distillation (MOPD) into GR to distill two single-objective teachers into a student at each token. Yet, we find that in vanilla MOPD, (i) teachers exhibit unequal distillation difficulty at each token (i.e., token-level imbalance), and (ii) the harder-to-distill teacher varies across tokens (i.e., cross-token inconsistency), leaving each objective poorly learned at certain steps and hindering positive item retrieval for both objectives, leading to suboptimal performance. To address this challenge, we propose KL-MOPD (KL-Weighted Multi-Objective On-Policy Distillation), a simple yet effective distillation method that allocates token-wise teacher weights proportional to the initial student-teacher token-wise KL divergences, thereby strengthening supervision for the harder-to-distill teacher at every decoding step. Despite its simplicity, we theoretically prove that KL-MOPD strictly outperforms vanilla MOPD in the additive joint recommendation objective under mild conditions. Moreover, KL-MOPD is plug-and-play and computationally efficient, requiring only a one-time pre-training weight calibration with negligible training overhead. Extensive experiments on three public benchmarks, a billion-scale industrial dataset, and online A/B tests on our e-commerce advertising platform demonstrate that KL-MOPD significantly improves both individual and joint recommendation performance. The anonymous code is provided for peer review.

cs.IR↗
arXiv2026

In the Driver's Seat: A Multi-Company Study on the Reality of Autonomous Driving System Testing

Autonomous driving systems (ADS) are increasingly deployed in complex real-world environments, creating strong demands for rigorous testing to establish their functionality and safety. However, ADS testing remains difficult because practitioners must determine not only what and how to test, but also whether the resulting evidence is sufficient to support progression and release decisions. To understand how this is handled in practice, we conducted semi-structured interviews with experts involved in ADS development and testing across nine companies in six countries. Through thematic analysis, we examined how practitioners organize ADS testing to produce and assess evidence, and what limits confidence in that evidence. Our findings show that industrial ADS testing can be understood as an evidence-building process involving three recurring activities: planning what evidence is needed, producing and evaluating evidence across complementary testing environments, and deciding when testing is sufficient. We further identified recurring limitations concerning scenario adequacy, simulation and deployment fidelity, test execution and observability, and assurance and decision sufficiency. Practitioners reported several approaches for addressing these limitations, although important gaps remain. Based on these findings, we propose an evidence-centered closed-loop testing framework that connects testing intent, scenario selection, evidence production, acceptance decisions, and operational feedback. The study provides an industry-grounded account of how ADS testing is organized, where confidence in current testing breaks down, and how testing processes can be structured more systematically.

cs.SE↗
arXiv2026

TwinRouterBench: Fast Static and Live Dynamic Evaluation for Realistic Agentic LLM Routing

LLM routing matters most in long-horizon applications such as coding agents, deep research systems, and computer-use agents, where a single user request triggers many model calls. Routing each call to the cheapest sufficient model can cut costs without sacrificing quality, yet existing router benchmarks evaluate routers only on one-shot prompts. They never expose the router-visible prefix at an intermediate agent step, never test whether a cheaper replacement preserves downstream task success, and often rely on online LLM judges at evaluation time. We introduce TwinRouterBench, a step-level routing benchmark with two tracks. The static track provides 970 router-visible prefixes from 520 instances across SWE-bench, BFCL, mtRAG, QMSum, and PinchBench, each paired with an execution-verified target tier estimated under a released downgrade-and-cascade protocol; scoring is deterministic arithmetic over tier labels, trajectory membership, and token costs, with no online evaluator-side LLM judge. The dynamic track supplies a harness that runs routers on the full 500-case SWE-bench Verified suite; in this paper we report a 100-case held-out evaluation disjoint from the static SWE supervision split. At each LLM call the router selects a concrete model from a locked pool, and success is measured by official task resolution and realized API spend. The two tracks support fast offline iteration followed by end-to-end validation under live agent execution. Code and data are available at https://github.com/CommonstackAI/TwinRouterBench.

cs.LG↗
arXiv2026

Perfect Born Sampling of Symmetric Thermal Tensor Network States

Accurate calculations of correlated fermions at low temperatures constitute a major challenge in quantum many-body physics, most notably in the studies of high-$T_c$ superconductivity and ultracold atom optical lattices. Stochastic sampling of tensor-network states offers a promising route to tackle this problem; however, existing schemes have long faced a fundamental dilemma---sampling efficiency and symmetry acceleration \textit{cannot} be achieved simultaneously. Here we propose a perfect Born sampling approach for thermal tensor networks, which performs importance sampling directly from the purified density matrix via the Born rule and incorporates Abelian and non-Abelian symmetries by sampling symmetry quantum numbers. We benchmark the method, realized as both stochastic matrix product states (stoMPS) and stochastic projected entangled pair states (stoPEPS), on large-scale quantum lattice models. Using stoMPS, we accurately simulate the square-lattice Hubbard model on cylinders up to width $W=10$, and study the triangular-lattice Hubbard model down to $T/t = 1/64$. We reveal scalar chiral order at half filling and kinetic ferromagnetism upon electron doping, with the latter directly compared to ultracold atom experiments. We further extend the stoMPS method to compute finite-temperature quantum dynamics, as demonstrated by the optical conductivity of the Hubbard model. We showcase the generalization of stoPEPS on up to $20\times20$ square-lattice quantum Ising model, where accurate results are obtained near its quantum critical point. Our method combines high sampling efficiency with full symmetry acceleration. The resulting state-of-the-art framework captures both equilibrium and dynamical properties down to ultralow temperatures, thereby providing a powerful theoretical tool for correlated quantum systems, particularly fermion Hubbard quantum gas microscopy.

cond-mat.str-el↗
arXiv2026

Machine Learning-Enabled Wireless Power Transfer: A Task-Oriented Review of Structural Optimization, State and Parameter Estimation, and Adaptive Control

Wireless power transfer (WPT) systems must maintain efficient and safe energy delivery under variations in geometry, coupling, load, alignment, and operating conditions. Machine-learning methods are increasingly applied to accelerate structural design, estimate operating parameters and states, and support adaptive control and scheduling. This review develops a task-oriented framework organized around three interconnected domains: structural design and optimization, parameter and state estimation, and control and optimization. Supervised learning, reinforcement learning, and hybrid or physics-informed methods are compared in terms of data requirements, dependence on physical models, information sources, computational requirements, accuracy, online response, application scope, and optimization objectives. Supervised learning provides rapid surrogate modeling, regression, and classification when representative labeled data are available. Reinforcement learning addresses sequential control and scheduling but introduces challenges in reward design, safe exploration, and simulation-to-reality transfer. Hybrid methods integrate physical knowledge with data-driven models to improve data efficiency and physical consistency. The principal barriers to engineering deployment include limited and biased datasets, weak out-of-distribution generalization, insufficient physical consistency, unsafe online adaptation, embedded-resource constraints, and the absence of standardized benchmarks. Future research should prioritize uncertainty-aware and physics-informed learning, safe simulation-to-reality transfer, lightweight edge deployment, distributed coordination, reproducible evaluation, and digital-twin-assisted development.

eess.SY↗
arXiv2026

ReBeCA: Causality-Inspired Behavior Analysis for Understanding the Self-Reflection of Language Models

Self-reflection is a basic self-improvement skill of language agents, yet it does not reliably improve their outputs, where the underlying behavioral mechanism remains opaque. We introduce **ReBeCA** (self-**Re**flection **Be**havior explained through **C**ausality-inspired **A**nalysis), a framework for identifying and validating outcome-relevant reflective behaviors from recorded trajectories. ReBeCA uses **reflective codes**, an inductive coding schema, to represent each trajectory as a sequence of reflective behaviors. It learns dependencies among these behaviors and the outcome with causal discovery, selects the behaviors that best predict the outcome of held-out trajectories, and tests them for consistency across data subsets and by prompt-based steering. In a controlled case study of four Qwen3 models under fixed-round Self-Refine on mathematical reasoning and translation, ReBeCA identifies compact, round-specific behavioral sets with superior held-out predictive fit and uses prompt-based steering on unseen problems to reveal how individual and combined behavioral directives affect performance (Holm-adjusted $p = .020$ on Qwen3-4B). ReBeCA offers a reusable framework that turns agent trajectories into interpretable, verifiable behavioral hypotheses without requiring model internals or preselecting the behaviors to test.

cs.CL↗
arXiv2026

InfraBench: Evaluating Infrastructure Agents Across Layers, Lifecycle, and Risk

Managing modern computing infrastructure has become a steadily harder problem due to the ever-increasing complexity. Recent advances in AI agents create a timely opportunity to automate infrastructure management tasks, but it remains unclear how well such agents can handle real-world infrastructure complexity. We present InfraBench, a benchmark for infrastructure agents organized around system layers, the operational lifecycle, and risk. The current prototype couples container, VM, and bare-metal backends with task-specific verifiers that distinguish functional repair from durability and cleanup, and analyzes retained action-trace findings for unsafe shortcuts. We evaluate 18 agent-model configurations on 12 seed tasks, with three attempts per task. Mean effective scores range from 45.5% to 89.2%; the strongest configuration completes 77.8% of attempts perfectly. In the available lifecycle checks, pass rates fall from 91.5% for functional repair to 80.9% for durability and 38.9% for cleanup. These results show why immediate recovery is an incomplete measure of operational capability. We use this evidence to motivate a planned expansion toward 40-60 tasks centered on managed cloud infrastructure and redesign after incidents, including provider-native Kubernetes networking. This extension is a research direction; the reported results remain those of the 12-task prototype. InfraBench, its tasks, evaluation harness, and live leaderboard are available at infraben.ch.

cs.AI↗
arXiv2026

Computing Multi-Slot First-Price Pacing Equilibria at Scale via Entropic Smoothing

Pacing equilibria describe how budget-constrained advertisers clear an ad market, but computing one at production scale is hard when every request fills several slots: the allocation jumps whenever two campaigns' scores cross. We consider the first-price pacing equilibrium (FPPE) of auctions with $K$ homogeneous slots and unit-demand campaigns, whose Eisenberg-Gale dual reduces, via a TopK identity, to a convex program over bids. Regularizing each request's allocation with the binary (Fermi-Dirac) entropy gives a smooth, strictly convex dual whose gradient is exactly spend minus budget and whose evaluation splits into independent one-dimensional threshold solves, one per request. The smoothed equilibrium converges to the hard one at an explicit $O(\sqrt{T})$ rate, and a duality gap certifies how far any bids are from the hard equilibrium. We implement an efficient multi-GPU solver and test it on a production ad workload. We use it to compute an equilibrium price anchor: each campaign's hindsight equilibrium cost per key result.

math.OC↗
arXiv2026

Towards channel foundation models (CFMs): Motivations, methodologies and opportunities

Artificial intelligence (AI) models for wireless systems are commonly optimized for fixed tasks, channel distributions, and system configurations, which limits their reuse when downstream objectives or operating conditions change. Channel foundation models (CFMs) address this limitation by pretraining reusable models on channel-informative observations and adapting their parameters or representations across channel-related tasks and conditions. However, existing studies differ substantially in their input modalities, pretraining targets, retained components, and transfer settings, making it difficult to compare how the methods are constructed, reused, and evaluated. This paper surveys CFM research through the central question of how channel-informative pretraining is converted into reusable downstream capabilities. We adopt a channel-centered scope covering explicit channel state information, received-signal observations, and auxiliary environmental or geometric context. Existing methods are organized according to the source of their learning targets into observation-space modeling, relational representation learning, latent-space prediction, task-target learning, system-utility optimization, and hybrid combinations. For representative methods, we compare how reusable parameters are learned and adapted, their downstream applications, and the settings used to evaluate transfer. The resulting framework connects CFM methodologies to physical-layer processing, radio access network and medium access control tasks, integrated sensing and communication, and radio-environment modeling. A matched case study further compares transfer for channel extrapolation, beam prediction, and positioning under common backbone and source-data settings. Finally, we discuss open problems in data construction, transfer evaluation, efficient adaptation, deployment, reliability, privacy, and security.

eess.SP↗