arXiv subjects

Hao Yin

Publications and source records attributed to Hao Yin.

At least 19 recordsLinked to original sources

arXiv2026

Data-Driven Brownian Reflection Control

We study a data-driven reflection control problem for a Brownian model with unknown drift and volatility. We first propose a learn-then-optimize (LTO) algorithm: it estimates the policy-relevant parameter during exploration, plugs the estimate into the optimality equation, and exploits the resulting policy---achieving an $O(\sqrt{T})$ finite-time expected regret bound. We further propose two algorithms, adaptive-updating (AU) and full-history adaptive-updating (AU-FH), which continuously update the estimator and reflecting level, attaining an improved $O(\log T)$ regret bound. Notably, AU-FH algorithms leverages all historical data, yielding better performance in numerical simulations. Our analysis decomposes regret into exploration, transient, and learning components. Transient regret from nonstationarity is bounded by the time-integrated deviation of the transition semigroup from stationarity evaluated on the holding cost, which can be further bounded via a Foster-Lyapunov inequality for exponential convergence of the controlled reflected Brownian motion (RBM). For learning regret, we establish local regularity properties together with consistency and mean-squared error bounds for the estimator, which control the stationary cost gap between the learned and optimal reflection policies. In addition, we leverage the monotonicity of the moving boundary Skorokhod map to derive moment bounds for the AU algorithms' switching states, via pathwise comparison with fixed-boundary RBMs.

math.OC↗

arXiv2026

Learning-Assisted Congestion-Aware Route Scheduling for Semiconductor Fab Material Control Systems

Automated material handling systems in semiconductor fabs are operated by a material control system (MCS) that must schedule a relay route for every transport command online, before execution. This is a data-driven scheduling problem in which route cost is dominated in the upper tail by queueing at heterogeneous, partially observable relay equipment, so route selection requires estimating both delivery time and congestion risk at the decision moment. This paper proposes a transport-network-aware dynamic congestion representation (TN-DCR). Built on a static directed transport graph induced by historically observed relay segments, TN-DCR combines structural route priors, multi-window network-wide congestion context, route-level bottleneck exposure, and an inductive graph-aware route embedding, all constructed under a prediction-time-safety invariant that admits only information observed strictly before the prediction moment. The representation feeds separate queue- and transfer-time regressors and an ordinal multi-label classifier producing calibrated multi-threshold exceedance scores, with an empirical-Bayes stock-key residual correction reducing systematic queue-time underprediction. The predictions serve as costs in a risk-constrained route-scheduling rule that minimizes predicted delivery time subject to a bound on extreme-congestion probability, embedding the learned predictors within a lightweight operations-research decision model. In a controlled closed-loop evaluation, mean delivery time falls by 16.4\% and internal resource waiting time by 22.6\% while throughput remains essentially unchanged.

cs.AI↗

arXiv2026

MyoMechanix: Biomechanically-Grounded Compositional Skilled Activity Understanding and Coaching

Existing action quality assessment (AQA) datasets and methods rely primarily on visual inputs such as RGB and pose, overlooking physiological dynamics such as muscle mechanics and often modeling actions as monolithic patterns. These limitations hinder fine-grained, biomechanically grounded feedback. We introduce MyoMechanix, a multimodal ecosystem for weight-loaded actions that aligns motion with muscle activity. Expert-annotated, it contains 7,500+ samples of 20 actions from 38 subjects, with synchronized multiview RGB video, 3D pose, sEMG, and additional physiological signals, forming the largest multimodal AQA benchmark to date. We further construct the Fitness Knowledge Graph (FKG), which organizes expert annotations into structured relationships among actions, phases, key steps, errors, and corrective feedback, enabling compositional scoring and interpretable assessment. Building on these representations, we develop CUBIST (Compositional Ontological Reasoning Engine), which performs decomposition-analysis-recomposition for fine-grained error attribution and feedback generation. We also establish MyoMechanix-AQA, MyoMechanix-VideoQA, and a novel MyoMechanix-Video2EMG task. Experiments show that multimodal sensing and structured representations improve performance, interpretability, and error attribution, with CUBIST achieving state-of-the-art results; VideoQA enhances language-grounded action understanding; and Video2EMG suggests video-based alternatives to costly EMG sensing. MyoMechanix advances skilled activity understanding toward biomechanically grounded, multimodal, and compositional reasoning for Physical AI applications in fitness, rehabilitation, healthcare, and machine learning. Project page: https://haoyin116.github.io/MyoMechanix/

cs.CV↗

arXiv2026

UBio-MolFM: Enabling Biomolecular Dynamics at DFT Accuracy and $10^5$ Atoms with One Untuned Potential

Ion conduction, membrane permeation and metal recognition hinge on electronic structure, yet first-principles simulation reaches only hundreds of atoms. UBio-MolFM lifts that ceiling: a foundation model trained on 160 million quantum-chemical labels, its receptive field spanning non-covalent distances at near-linear cost. The barrier is cost, not principle. One untuned potential keeps force error near 20 meV/{\AA} past a thousand atoms, reproduces water's X-ray structure and ion hydration, and holds an RNA Mg$^{2+}$ site without ion-specific parameters. Cyclosporine A pays 3.5 kcal/mol in water for its permeable conformer, gated by one kinetically asymmetric hydrogen bond that a fixed-charge model flattens. In a 108,964-atom KcsA channel on one GPU, the relaxed four-ion column is anhydrous in all five replicas, in direct contact in four---the knock-on geometry ten fixed-charge simulations never form. It remains orders of magnitude costlier. Where electronic structure decides the answer, first-principles simulation is in reach.

physics.chem-ph↗

arXiv2026

Learning from Unreachable Rewards: Hint-Conditioned Reinforcement Learning for Generative Recommendation

Semantic-ID generative recommenders represent each item as a short sequence of discrete semantic tokens and predict the next item by autoregressively generating this token sequence. This paradigm enables a unified generation interface for item IDs, histories, and item text, but it also creates a structured optimization bottleneck during reward-based post-training: when an early semantic token enters the wrong branch of the item-token space, finite rollout groups rarely reach the ground-truth item, so group-relative optimization receives identical zero rewards and produces no useful advantage. We propose Hint-Conditioned Generative Recommendation (HCGRec), a semantic-ID generative recommendation framework that recovers learning signal for such hard training instances. HCGRec diagnoses each instance with checkpoint rollouts and supplies a minimal target-prefix hint only when the current generator cannot reach the correct item. The model then generates the unhinted suffix under the hinted semantic branch, turning zero-reward groups into informative comparisons over item-token completions. Hinting also changes token identity: hinted prefix tokens are oracle-provided item context, while unhinted suffix tokens are sampled generation actions. We therefore introduce hint-aware credit decomposition, using supervised learning to preserve item-semantic and prefix-structure alignment for hinted tokens and GRPO to optimize the sampled suffix. Experiments on sequential recommendation benchmarks show that HCGRec substantially improves over supervised fine-tuning and vanilla reward-based post-training, while reducing zero-advantage training samples from over 70% to below 20%. The code is accessible at https://github.com/WncFht/GRec.

cs.IR↗

arXiv2026

A Multi-Agent Framework for Zero-Dimensional Reduced-Order Model Planning

Zero-dimensional reduced-order models (0D ROMs) are central to multi-dimensional design workflows for high-end complex equipment. However, the planning process currently relies on manual expertise, limiting topological exploration and prolonging iterations. Even traditional optimization methods such as Genetic Algorithms (GA) are typically confined to local parameter tuning. Although Large Language Model (LLM) agents have shown promise in exploring large sample spaces, and frameworks such as Chain of Thought (CoT) and Reason and Act (ReAct) improve reasoning reliability, while Retrieval-Augmented Generation (RAG) overcomes domain knowledge barriers, a single agent still falls short for the long-horizon and highly coupled nature of complex 0D ROM planning. This paper proposes the Zero-dimensional reduced-order model CO-Planning framework (Z-COPA), a multi-agent architecture featuring a Symbolic Action Graph Engine (SAGE) and a MILP-Guided Navigation (MGN) optimizer. Its core innovation is a dedicated graph representation method that accurately encodes the 0D flow network topology, converting the empirical planning process into a rigorous graph structure optimization problem. We validate the forward and inverse design capabilities and generalization performance of Z-COPA on two real aircraft engine secondary-air systems, two IEEE power-distribution reconfiguration benchmarks, and two water-distribution network benchmarks. The results show superior task completion quality, obtaining the best performance in both forward and reverse design of air systems. Z-COPA disrupts the traditional 0D model planning paradigm, providing a new technical approach for exploring broader topological space and achieving highly automated, globally optimal air system architectures.

cs.LG↗

arXiv2026

On the ends of Willmore surfaces with curvature decay

For a properly embedded Willmore surface $\Sigma$ in $\mathbb R^3$, we prove that if the scale-invariant second fundamental form is sufficiently small near infinity, the surface has finitely many ends. Moreover, if this scale-invariant quantity vanishes at infinity, or if there is only one end, the total $L^2$-norm of the second fundamental form is finite.

math.DG↗

arXiv2026

OpenClawBench: Benchmarking Process-side Anomalies in Real-world Agent Execution Trajectories

Task success can hide process anomalies in real-world agent executions. An agent may pass the final task oracle while still accumulating unresolved ambiguity, unsafe external writes, ignored errors, weakly grounded commitments, or capability-boundary overcommitment. We study this mismatch as the Outcome-Process Gap and introduce OpenClawBench, a large-scale dataset for measuring and supervising process-side anomalies in real agent execution processes. OpenClawBench is built from BFCL-driven OpenClaw sessions produced by 6 source models and contains 31,264 annotated trajectories. It aligns task-oracle outcomes with structured process evidence. FullTax converts the aligned trajectories into structured anomaly supervision: binary labels, supporting evidence, onset/span localization, severity, recoverability, and a 5-class anomaly taxonomy. Using OpenClawBench, we make the Outcome-Process Gap measurable. Among 31,135 oracle-passing executions, 2,904 are still labeled process-anomalous under FullTax. These results show that success-only evaluation misses a concrete class of process-side failures in real agent executions. A LoRA-fine-tuned Gemma 3 12B detector trained on the high-confidence FullTax supervised pool reaches binary F1=0.729 on the cleaner-labels held-out test split. Together, OpenClawBench turns real agent execution logs into auditable and reusable supervision for studying, diagnosing, and operationally monitoring runtime agent reliability.

cs.AI↗

arXiv2026

Discovery of unobservable parameters via physical embedding

Recovering a source signal from indirect measurements often requires estimating latent parameters, such as wireless channel states or MRI coil sensitivities, that cannot be directly observed. Here, we introduce Physics-Embedded Inverse Learning (PEIL), in which a learned estimator predicts these parameters and a fixed, physics-based inverse operator uses them to reconstruct the signal, so that training requires only the source signal as supervision. In systems where multiple parameter combinations can reconstruct the signal equally well, the estimator exploits this freedom to coordinate parameters that compensate for residual modelling errors rather than match ground-truth parameters. In high-mobility wireless communications, PEIL discovers task-optimal configurations that outperform baselines given access to ground-truth parameters, enabling zero-shot generalisation and over 20-fold reduction in training data relative to supervised baselines. To test whether these properties extend across physical domains, we apply PEIL to parallel MRI, where it discovers physically interpretable coil sensitivity maps without calibration scans, yielding reconstructions grounded purely in acquired measurements. These results demonstrate that non-identifiability, conventionally a liability, becomes a resource when the learning objective targets reconstruction quality rather than parameter accuracy.

eess.SP↗

arXiv2026

Circular polarization images of Sgr A* for different magnetic field geometries

Sgr A* exhibits a persistent negative circular polarization (CP) at 230\,GHz, offering a powerful probe of the magnetic field geometry in its accretion flow. Using a stationary semi-analytic radiatively inefficient accretion flow (RIAF) model in Kerr spacetime with polarized radiative transfer, we systematically analyze CP images for six poloidal magnetic field configurations across varying black hole spins, inclinations, and field polarities. We find that CP production is dominated by Faraday conversion in radial, parabolic, quadrupole, and combined geometries, but by intrinsic emission in dipole and vertical fields. The radial and parabolic configurations produce the polarity-invariant net CP, while dipole and vertical fields yield the polarity-sensitive one. As the accretion disk is prograde with respect to the black hole spin, the CP production across all six field geometries is found to be lower at high spin case, while the situation is more complicated in the retrograde case. Moreover, the net CP observed from edge-on views $V_{\rm net} \approx 0$ except for the quadrupole geometry. Comparing with ALMA data, the reversed-field model is excluded at high inclinations and then the magnetic field geometry of Sgr A* is constrained.

astro-ph.HE↗

arXiv2026

The Yang-Mills equation near instanton-anti-instanton configurations

We study the question of whether a sequence of non-instanton Yang-Mills connections can limit to a bubbling configuration composed only of instantons. In the case that the Uhlenbeck limit and the bubbles are of opposite charge, we determine an obstruction coming from deformations of the Uhlenbeck limit. As an application, we prove that instantons are the only solutions of the $\mathrm{SU}(2)$ Yang-Mills equation on $\mathbb{R}^4$ with energy less than $4\pi^2 \left( |\kappa| + 2 \right) + \varepsilon_\kappa,$ where $\kappa$ is the charge. We also prove discreteness of the energy spectrum on the trivial $\mathrm{SU}(2)$-bundle in the range $\left[ 0, 16 \pi^2 \right).$

math.DG↗

arXiv2026

On the blow-up of harmonic maps from surfaces to homogeneous manifolds

We study harmonic map sequences from surfaces to compact homogeneous spaces. For sequences developing a single bubble, we derive refined asymptotic expansions in the neck region and prove new obstruction relations among the leading coefficients. These strengthen earlier results by converting an inequality into an equality. For weakly conformal maps, this yields geometric constraints: in low dimensions the tangent planes of the limit map and bubble must coincide, while in higher dimensions they are isoclinic.

math.DG↗

arXiv2026

Subcycle videography of lightwave-driven Landau-Zener-Majorana transitions in graphene

Strong light fields have unlocked previously unthinkable possibilities to tailor coherent electron trajectories, engineer band structures and shape emergent phases of matter all-optically. Unravelling the underlying quantum mechanisms requires a visualisation of the lightwave-driven electron motion directly in the band structure. While photoelectron momentum microscopy has imaged optically excited electrons averaged over many cycles of light, actual subcycle band-structure videography has been limited to small electron momenta. Yet lightwave-driven elementary processes in quantum materials often occur throughout momentum space. Here, we introduce attosecond-precision, subcycle band-structure videography covering the entire first Brillouin zone (BZ) and visualize one of the most fundamental but notoriously elusive strong-field processes: non-adiabatic Landau-Zener-Majorana (LZM) tunnelling. The interplay of field-driven acceleration within the Dirac-like band structure of graphene and periodic LZM interband tunnelling manifest in a coherent displacement and distortion of the momentum distribution at the BZ edge. The extremely non-thermal electron distributions also allow us to disentangle competing scattering processes and assess their impact on coherent electronic control through electron redistribution and thermalization. Our panoramic view of strong-field-driven electron motion in quantum materials lays the foundation for a microscopic understanding of some of the most discussed light-driven phenomena in condensed matter physics.

cond-mat.mes-hall↗

arXiv2026

Video-OPD: Efficient Post-Training of Multimodal Large Language Models for Temporal Video Grounding via On-Policy Distillation

Reinforcement learning has emerged as a principled post-training paradigm for Temporal Video Grounding (TVG) due to its on-policy optimization, yet existing GRPO-based methods remain fundamentally constrained by sparse reward signals and substantial computational overhead. We propose Video-OPD, an efficient post-training framework for TVG inspired by recent advances in on-policy distillation. Video-OPD optimizes trajectories sampled directly from the current policy, thereby preserving alignment between training and inference distributions, while a frontier teacher supplies dense, token-level supervision via a reverse KL divergence objective. This formulation preserves the on-policy property critical for mitigating distributional shift, while converting sparse, episode-level feedback into fine-grained, step-wise learning signals. Building on Video-OPD, we introduce Teacher-Validated Disagreement Focusing (TVDF), a lightweight training curriculum that iteratively prioritizes trajectories that are both teacher-reliable and maximally informative for the student, thereby improving training efficiency. Empirical results demonstrate that Video-OPD consistently outperforms GRPO while achieving substantially faster convergence and lower computational cost, establishing on-policy distillation as an effective alternative to conventional reinforcement learning for TVG.

cs.CV↗

arXiv2025

Hyperion: Low-Latency Ultra-HD Video Analytics via Collaborative Vision Transformer Inference

Recent advancements in array-camera videography enable real-time capturing of ultra-high-definition (Ultra-HD) videos, providing rich visual information in a large field of view. However, promptly processing such data using state-of-the-art transformer-based vision foundation models faces significant computational overhead in on-device computing or transmission overhead in cloud computing. In this paper, we present Hyperion, the first cloud-device collaborative framework that enables low-latency inference on Ultra-HD vision data using off-the-shelf vision transformers over dynamic networks. Hyperion addresses the computational and transmission bottleneck of Ultra-HD vision transformers by exploiting the intrinsic property in vision Transformer models. Specifically, Hyperion integrates a collaboration-aware importance scorer that identifies critical regions at the patch level, a dynamic scheduler that adaptively adjusts patch transmission quality to balance latency and accuracy under dynamic network conditions, and a weighted ensembler that fuses edge and cloud results to improve accuracy. Experimental results demonstrate that Hyperion enhances frame processing rate by up to 1.61 times and improves the accuracy by up to 20.2% when compared with state-of-the-art baselines under various network environments.

cs.DC↗

arXiv2025

REVISOR: Beyond Textual Reflection, Towards Multimodal Introspective Reasoning in Long-Form Video Understanding

Self-reflection mechanisms that rely on purely text-based rethinking processes perform well in most multimodal tasks. However, when directly applied to long-form video understanding scenarios, they exhibit clear limitations. The fundamental reasons for this lie in two points: (1)long-form video understanding involves richer and more dynamic visual input, meaning rethinking only the text information is insufficient and necessitates a further rethinking process specifically targeting visual information; (2) purely text-based reflection mechanisms lack cross-modal interaction capabilities, preventing them from fully integrating visual information during reflection. Motivated by these insights, we propose REVISOR (REflective VIsual Segment Oriented Reasoning), a novel framework for tool-augmented multimodal reflection. REVISOR enables MLLMs to collaboratively construct introspective reflection processes across textual and visual modalities, significantly enhancing their reasoning capability for long-form video understanding. To ensure that REVISOR can learn to accurately review video segments highly relevant to the question during reinforcement learning, we designed the Dual Attribution Decoupled Reward (DADR) mechanism. Integrated into the GRPO training strategy, this mechanism enforces causal alignment between the model's reasoning and the selected video evidence. Notably, the REVISOR framework significantly enhances long-form video understanding capability of MLLMs without requiring supplementary supervised fine-tuning or external models, achieving impressive results on four benchmarks including VideoMME, LongVideoBench, MLVU, and LVBench.

cs.CV↗

arXiv2025

WEST: LLM based Speech Toolkit for Speech Understanding, Generation, and Interaction

In this paper, we present WEST(WE Speech Toolkit), a speech toolkit based on a large language model (LLM) for speech understanding, generation, and interaction. There are three key features of WEST: 1) Fully LLM-based: Standing on the shoulders of giants by reusing mature architectures, ecosystems (e.g., Hugging Face), and methods (e.g., sequence packing) from large models. 2) Full-stack: Supports tasks such as recognition, synthesis, understanding, dialogue, and multimodal capabilities, with extensibility to incorporate open-source models. 3) Simple and Stupid: A simple and stupid speech toolkit that everyone can Touch. In addition, WEST provides two types of recipes, models, and experimental results. The first is entirely based on open-source models and open-source data, allowing users to fully reproduce the experiments in this paper and serving as a verification system or minimal system baseline. The second is trained on massive data, offering superior performance so the user can directly apply it out of the box. WEST is publicly avilable at https://github.com/wenet-e2e/west/

cs.CL↗

arXiv2025

Cross-modal Full-mode Fine-grained Alignment for Text-to-Image Person Retrieval

Text-to-Image Person Retrieval (TIPR) is a cross-modal matching task designed to identify the person images that best correspond to a given textual description. The key difficulty in TIPR is to realize robust correspondence between the textual and visual modalities within a unified latent representation space. To address this challenge, prior approaches incorporate attention mechanisms for implicit cross-modal local alignment. However, they lack the ability to verify whether all local features are correctly aligned. Moreover, existing methods tend to emphasize the utilization of hard negative samples during model optimization to strengthen discrimination between positive and negative pairs, often neglecting incorrectly matched positive pairs. To mitigate these problems, we propose FMFA, a cross-modal Full-Mode Fine-grained Alignment framework, which enhances global matching through explicit fine-grained alignment and existing implicit relational reasoning -- hence the term ``full-mode'' -- without introducing extra supervisory signals. In particular, we propose an Adaptive Similarity Distribution Matching (A-SDM) module to rectify unmatched positive sample pairs. A-SDM adaptively pulls the unmatched positive pairs closer in the joint embedding space, thereby achieving more precise global alignment. Additionally, we introduce an Explicit Fine-grained Alignment (EFA) module, which makes up for the lack of verification capability of implicit relational reasoning. EFA strengthens explicit cross-modal fine-grained interactions by sparsifying the similarity matrix and employs a hard coding method for local alignment. We evaluate our method on three public datasets, where it attains state-of-the-art results among all global matching methods. The code for our method is publicly accessible at https://github.com/yinhao1102/FMFA.

cs.CV↗