Searcharxiv⌕ Search

arXiv subjects

Haoyuan Li

Publications and source records attributed to Haoyuan Li.

At least 19 recordsLinked to original sources

Trait Collapse in Text-to-Image Generation: Evaluating and Mitigating Personality Inconsistency in Visual Storytelling

In visual storytelling, a believable character must not only look consistent across scenes but also \textit{behave} in ways that reflect their personality. Current text-to-image (T2I) models preserve visual appearance effectively, yet consistently fail to maintain personality-relevant behavior---a failure we term \textit{trait collapse}. We show that conventional metrics (CLIP Score, DINO similarity) are blind to this failure, and that abstract personality descriptors and scene-level emotion labels do not produce consistent improvements over neutral prompts across models. To systematically diagnose trait collapse, we introduce \textbf{ViStoryBench-Personality}, a story-level benchmark combining an expert-validated Visual Cues Lexicon (Fleiss' $κ= 0.757$) with a six-dimensional metric suite covering personality alignment, cross-scene stability, and behavioral coherence. Evaluating seven T2I systems, we confirm a persistent semantic gap between abstract trait labels and visual evidence, and find that visual identity consistency does not imply personality consistency. To bridge this gap, we propose \textbf{Trait-Behavior-Visual (TBV)}, a training-free prompting framework that decomposes personality traits into renderable visual constraints. TBV improves Personality Alignment Score by up to $+22.1\%$, with larger gains in models with stronger language-conditioning capacity.

cs.CV↗

A semigroup approach for rough partial differential equations with jumps

We establish an infinite-dimensional sewing lemma for discontinuous paths on a scale of Banach spaces. This allows us to construct Young and rough convolutions with jumps. Based on this we analyze the local and global well-posedness of semilinear parabolic rough partial differential equations (PDEs) with jumps in mixed space-time regularity spaces.

math.AP↗

Beyond Federated Learning: Decentralized Agentic Systems for Self-Evolving Clinical Intelligence

Clinical expertise is inherently distributed: it resides not in any single dataset or institution but in the accumulated reasoning experiences of thousands of clinicians navigating diverse patient populations, resource constraints, and evolving treatment protocols. Yet current approaches to scaling medical AI, from centralized large language models to federated learning, remain anchored to a data-centric paradigm that treats knowledge as static parameters to be aggregated. In this Perspective, we argue for a fundamental shift from data aggregation to experience aggregation. We propose Decentralized Agentic Systems (DAS), a framework in which distributed AI agents operating across healthcare institutions synthesize collective clinical intelligence by sharing abstracted reasoning trajectories rather than raw data or model parameters. Unlike federated learning, which averages numerical gradients across homogeneous model architectures, DAS exchanges structured records of clinical reasoning, specifically the sequence of observations, hypotheses, tool invocations, and outcome feedback that constitute diagnostic experience. We examine the architectural principles, deployment considerations, governance requirements, and critical open challenges of this framework, laying the groundwork for a collaborative medical AI ecosystem where distributed expertise autonomously adapts to the complexities of longitudinal patient care.

cs.AI↗

Value-Rich, Differently Organized: Measuring Value Expression and Organization in AI-Generated Debates

Current evaluation of LLM-generated debate largely focuses on argumentative quality, leaving the value structure of generated arguments underexamined. We introduce a diagnostic metric framework for comparing how values are expressed and organized in human-reference and AI-generated debates. The framework separates value expression, the range of recognizable values a debate invokes, from value organization, how those values are distributed and sustained across turns. Grounded in Schwartz's Basic-10 value theory, it represents debates as multi-level value distributions over argumentative units, utterances, stances, and full debates, enabling metrics for value coverage, alignment, diffusion, and cross-turn dynamics. Across multiple generation systems, AI-generated debates are broadly value-expressive, often activating as many or more value dimensions than matched human references, but they organize these values differently. Their utterances are more locally value-diffuse, and source-blind judges distinguish value clarity from value focus. These findings show that debate evaluation should consider not only value presence, but also value organization. Code and supplementary materials are available at https://github.com/Maoyuan-li/Value-Expression-and-Organization.

cs.AI↗

Can We Believe AI Peer Review?

As large language models become widely used in scientific review and manuscript revision, AI scores are increasingly serving as feedback for assessing paper quality and guiding edits. Yet whether a single AI score reliably reflects the reviewer's judgment, and whether revision gains based on that score are endorsed by other reviewers, remain understudied. Using 200 papers from ICLR 2025 and 2026 and approximately 14,000 AI reviews, we study score reliability, score-distribution estimation, and cross-model agreement on revision gains. Repeated reviews fluctuate even under identical conditions. Finer-grained scoring protocols reduce score variance but also systematically shift mean scores, while the gains assigned to manuscript revisions depend strongly on the reviewer.

cs.AI↗

GeoThinker-$F$: Endogenous Predictive World Representations for Fast Spatial Intelligence

Existing geometry-aware Multimodal Large Language Models (MLLMs) increasingly rely on external 3D encoders to provide structural priors, but executing these additional encoder paths at inference introduces substantial latency and computational overhead. We propose GeoThinker-$F$, an efficient framework that internalizes spatiotemporal perception by transforming causal MLLM hidden states into predictive latent world representations. Our key insight is that intermediate semantic representations contain sufficient information to predict useful representations of current geometry and future semantics, without requiring an exact reproduction of features from a dedicated 3D encoder. To realize this idea, GeoThinker-$F$ adopts a decoupled two-stage training strategy. In the first stage, the MLLM backbone is frozen while lightweight Latent Frame Prediction (LFP) heads learn from an offline 3D geometry expert and next-frame semantic targets. In the second stage, the external expert is removed, and the model is jointly optimized to generate and selectively integrate its self-predicted priors through a specialized Spatial-Grounded Temporal (SGT) Fusion module. Extensive evaluations show that GeoThinker-$F$ eliminates runtime dependence on external 3D geometry encoders while maintaining competitive performance across diverse spatial intelligence benchmarks. It achieves up to a $3.6\times$ inference speedup, reducing latency from $1.45\,\mathrm{s}$ to $0.40\,\mathrm{s}$ while retaining strong spatial reasoning performance.

cs.CV↗

DynThinker: Unifying Static 3D and Dynamic 4D Spatial Reasoning

Achieving general spatial intelligence requires a unified understanding of both static 3D and dynamic 4D. However, we find that existing models struggle to excel at both: 3D post-training can weaken dynamic understanding, while 4D-specialized models often lack strong static 3D reasoning. Our analysis reveals that this spatial--dynamic interference is highly non-uniform across parameter groups, with the balance between static and dynamic capabilities particularly sensitive to changes in late-layer MLPs. To address this interference, we propose DynThinker, a unified framework for jointly strengthening static and dynamic spatial reasoning. Decoupled Exponential Moving Average (D-EMA) applies configurable group-wise averaging to constrain drift in interference-sensitive pretrained parameters while preserving flexibility for capability enhancement. Spatial-Dynamic Fusion (SDF) complements D-EMA by selectively integrating spatial and temporal priors from a dynamic geometry encoder into intermediate VLM representations under semantic guidance. Across ten metrics spanning nine static 3D and dynamic 4D benchmarks, DynThinker achieves the highest average score among all evaluated models. Notably, DynThinker achieves 56.0 on ReVSI, 67.2 on DSR-Bench, and 51.2 and 40.1 under the sample-wise and group-wise evaluation protocols of DSI-Bench, respectively, setting new state-of-the-art results among the compared open-source models across all four evaluations within a unified model.

cs.CV↗

Not Every Cue Is Evidence: Cue Supportability for Forensic Distillation in Face Forgery Detection

Explainable face forgery detection requires not only correct judgments but also forensic evidence supported by the input image. However, training cues are typically derived from the forgery generation pipeline. Since images available to the model are of limited quality and resolution, it is difficult to observe the fine-grained traces these cues refer to. Training methods that further supply such cues as privileged supervision exacerbate the problem, as the student is taught to articulate evidence it cannot perceive. The student consequently resorts to templated or over-generalized expressions, giving rise to the correct-answer, wrong-evidence (CAWE) failure. We propose a dual supportability audit that filters privileged cues before distillation. Each candidate cue must be predictable from the remaining privileged context and causally tied to high-attribution visual regions in the current image. Only cues satisfying both conditions receive substantial distillation weight, preventing unsupported details from leaking into the student. Experiments across eight benchmarks show that our method maintains or improves detection accuracy while significantly raising explanation faithfulness compared to existing paradigms.

cs.CV↗

PhysCodeGym: Benchmarking Coding Agents for Physically Verifiable 4D Reconstruction

Building physical simulations from video requires coding agents to translate observations into executable models and refine them through tool use. We introduce PhysCodeGym, a benchmark for physically verifiable 4D reconstruction with $156$ tasks covering rigid-body, liquid, deformable, and wind-driven motion. Agents autonomously inspect videos, write and run Blender programs, and revise their scenes using simulation feedback. Six metrics combine rule-based measurements with rubric-based judgments to assess visual and physical reconstruction quality. Program inspection, scene measurements, and counterfactual interventions test physical consistency under original and modified conditions. Experiments with frontier coding agents reveal a tradeoff between artifact quality and completion across reasoning-effort settings, while harness comparisons and agent traces link verification behavior with reconstruction performance. Case studies show that matching visible objects and events can coexist with physical defects. Counterfactual tests further expose latent contact and support failures in reconstructions that pass ordinary physical checks, highlighting the gap between reproducing observations and building reliable physical models. Code is available at https://github.com/ShaoShuai0605/PhysCodeGym.

cs.AI↗

Chinese Competitive Debating Dataset and Benchmark

Debate adjudication requires tracking how arguments develop through interaction, yet existing datasets rarely combine fine-grained debate transcripts with professional judgments collected during real competitions under a shared rubric. We introduce a dataset and benchmark for evaluating large language models' understanding of competitive Chinese-language debate at the match, stage, and speaker levels. We organized 182 matches and recruited 120 professional judges, with each match independently adjudicated by three judges using a predefined rubric. After excluding matches with incomplete records, the dataset contains 148 matches, 2,698 stages, and 20,542 exchange units, with manually verified transcripts and segmentation. It preserves original stage scores, match votes, best-debater ballots, and adjudication rationales. We define three tasks: winner-tendency prediction, stage-score prediction, and best-debater prediction. Zero-shot evaluation of multiple large language models yields a highest winner-prediction accuracy of 66.2%, a highest Pearson correlation of 0.250 between model stage scores and mean human ratings, and a highest best-debater prediction accuracy of 56.8%. The dataset and benchmark provide a testbed for studying large language models' understanding of interactive argumentation and their agreement with professional judges.

cs.CL↗

MCRI: A Four-Dimensional Framework for Analyzing and Evaluating Agent Skills

As agents evolve from single-tool systems into modular, composite architectures, skills are becoming an important mechanism for capability development and distribution. However, the academic community lacks a structured framework for systematically analyzing and evaluating skills. Drawing on information gain and behavioral constraint, we propose the four-dimensional MCRI Framework and operationalize it as MCRI-Eval, a large language model-based evaluation method. We evaluate MCRI-Eval using 63,812 public skills from the OpenClaw skill Hub, with 58,275 skill-conditioned model executions across BigCodeBench, BFCL-Fundamental, and Mind2Web. MCRI-Eval scores are positively associated with community popularity signals and achieve the highest downstream ranking agreement among the evaluated methods. MCRI-Eval also improves top-1 skill selection across all three benchmarks: compared with the strongest baseline on each benchmark, the skills selected by MCRI-Eval advance by 17.7, 22.8, and 19.6 percentile points in downstream performance rank on BigCodeBench, BFCL-Fundamental, and Mind2Web, respectively. These results indicate that MCRI-Eval provides a useful pre-execution signal for prioritizing promising skills before costly execution-based evaluation.

cs.AI↗

VR-JEPA: Learning Contrastive-State Latent Guidance for Generation-based Video Reasoning

Reasoning through video generation offers a promising path toward visual intelligence by modeling latent visual states and their dynamics. However, current video generation models often lack explicit guidance on how these states should evolve, leaving generated trajectories prone to physical and structural inconsistencies that undermine reasoning reliability. While the Video Joint-Embedding Predictive Architecture (V-JEPA) provides rich spatiotemporal priors learned through latent prediction, these general priors do not naturally adapt to the logical reasoning capabilities required for complex visual tasks. To bridge this gap, we propose VR-JEPA, a framework that aligns the V-JEPA predictor with task-specific reasoning logic through localized contrastive-state learning and uses its predicted latent trajectories to guide video generation for visual reasoning. Specifically, (i) we pair successful trajectories with generated alternatives under the same input conditions and use discrepancies in their V-JEPA representations to identify informative states and tokens for localized contrastive supervision. (ii) We further equip the V-JEPA predictor with skill-specific experts trained on anchor-task data, allowing the model to adaptively specialize its shared spatiotemporal priors across diverse cognitive domains. Together with skill-specific experts, this contrastive supervision enables VR-JEPA to predict latent trajectories that provide task-specific logical guidance for video generation. Comprehensive experiments on the large-scale VBVR-Pro-Bench dataset demonstrate that VR-JEPA achieves an $11.33\%$ relative improvement over the cutting-edge generation-based reasoning baseline, significantly mitigating physical artifacts and enhancing logical consistency.

cs.CV↗

Slow Thermalization and Long-Lived Coherence across Acoustic Phonon Branches in BAs

The exceptionally high thermal conductivity of cubic boron arsenide (BAs) arises from suppressed three-phonon scattering associated with its large acoustic--optical gap. In this regime, four-phonon processes become non-negligible, creating an unusual scattering hierarchy with unexplored consequences for nonequilibrium phonon dynamics. Here, time-resolved x-ray diffuse scattering reveals exceptionally slow, branch-dependent acoustic phonon thermalization and long-lived coherence. Exploiting branch sensitivity in one-phonon diffuse scattering, we resolve transverse acoustic (TA) and longitudinal acoustic (LA) thermalization times of 29.2 and 13.9~ps, respectively; even the LA timescale is at least three times that in common semiconductors. Phonon dispersion calculations assign coherent oscillations at 0.19 and 0.27~THz to the TA and LA modes, respectively; both persist with little decay over 50~ps, with comparable coherence found only in diamond. These findings provide a direct dynamical manifestation of the weak phonon scattering underlying the exceptional thermal transport of BAs.

cond-mat.mtrl-sci↗

Inter-dimension Dependence for Multi-Dimensional Evaluation of Open-Ended Text

LLM-as-a-judge methods are widely used for evaluating the quality of generated open-ended text. Such evaluations are generally multi-dimensional, since the error patterns in texts can be different for different dimensions. Therefore, reliable LLM judges should evaluate each target dimension independently. To quantify the extent to which LLM judges depend on non-target dimensions when evaluating a target dimension, i.e., inter-dimension dependence, we propose CorrGap. To measure this, CorrGap uses the difference in correlations between LLM-predicted scores and ground truth scores across different groups of texts. Using CorrGap, we show that inter-dimension dependence is pervasive across LLM judges in open-ended text evaluation tasks. To mitigate inter-dimension dependence, we propose DimCheck, a method that iteratively removes unrelated evidence from COTs generated by LLM judges in a step-wise way. We show that DimCheck mitigates inter-dimension dependence and outperforms strong baselines across three LLMs and four tasks. We also show that smaller trained LLMs can approximate larger LLMs in DimCheck, with much lower inference costs.

cs.CL↗

Propagation-Window Bounds in Degenerate Plant-Consumer Reaction-Diffusion Systems

We study traveling waves in high-dimensional plant--consumer reaction--diffusion systems with $N$ competing plants and $N$ associated consumers. We focus on the degenerate case in which consumer growth vanishes when plant populations are zero, so the standard Fisher--KPP linearization does not determine the leading-edge behavior. Under a weak-interaction condition, we identify a plant-driven lower threshold $s_{\mathrm P}$, a constructive upper threshold $s_{\mathrm C}^{\mathrm{ex}}$, and a universal necessary upper threshold $s_{\mathrm C}^{\mathrm{nec}}$, and prove $(s_{\mathrm P},s_{\mathrm C}^{\mathrm{ex}})\subseteq \mathcal S\subseteq[s_{\mathrm P},s_{\mathrm C}^{\mathrm{nec}})$, where $\mathcal S$ is the set of speeds admitting positive extinction-to-coexistence waves. Existence follows from new lower solutions that remove previously imposed diffusion and compatibility restrictions. Nonexistence below $s_{\mathrm P}$ follows from a Sturm argument, while center-manifold analysis and a Riccati crossing argument yield the finite upper-speed obstruction. An explicit two-species wave beyond $s_{\mathrm C}^{\mathrm{ex}}$ shows that this constructive threshold is not the true maximal speed.

math.AP↗

PanoWorld: Real-World Panoramic Generation

In this work, we aim to address the challenge of long-range memory in panoramic world models by exploiting the rotation-equivariant property of omnidirectional representations, where rotation can be treated as an implicit geometric transformation.Building on this insight, we propose PanoWorld, which simplifies camera trajectories into translations via fixed headings for both current-action modeling and long-range memory through Dense Panoramic Ray-Conditioning (DPRC) and Geometry-aware Memory Augmentation (GMA).Then, a three-stage training pipeline is introduced to progressively optimize each component. To better evaluate physical consistency under large-scale spatial variations and diverse illumination conditions, where existing datasets are relatively stable, we construct World360, a large-scale dataset consisting of both real-world video clips collected via panoramic unmanned aerial vehicles and high-quality simulated clips generated by AirSim360.Extensive experiments on World360 demonstrate the effectiveness of PanoWorld, outperforming alternative methods by a large margin.Our models, training code, and dataset will be publicly available. More information can be found on our project page: https://lihaoy-ux.github.io/panoworld-page/.

cs.CV↗

Skill-3D: Evolving Scene-Aware Skills for Agentic 3D Spatial Reasoning

This paper explores agentic 3D spatial understanding, i.e., MLLM agents performing 3D reasoning through tool use. Existing methods often misuse tools and exhibit biased tool preferences under 3D scenarios, leaving the agentic paradigm with only marginal gains over non-agentic strategies. We reveal that 3D spatial reasoning tasks are heterogeneous across scenes, while these agents apply a uniform tool-use strategy to all scenes rather than selecting tools according to the specific scene and task. To address this, we propose Skill-3D, a framework that learns self-evolving scene-aware skills. Specifically, Skill-3D identifies the task scene and records the agent's tool-use trajectory into a Scene Memory, where successful trajectories from similar scenes are aggregated and distilled into a reusable scene-aware skill, with failed ones attached to the skill as lessons. During training, once a similar scene recurs, the corresponding skill is injected to guide the agent, producing new trajectories whose successes and failures further refine the skill, forming a loop in which the memory and the skill library co-evolve. Experiments show that Skill-3D substantially improves tool utilization in 3D spatial reasoning (from 39% to 78% on VSI-Bench), driving the agent toward correct and sufficient tool use. For instance, it improves Gemini-3-Flash by 67% on MMSI-Bench. Furthermore, we conduct agentic post-training over skill-guided trajectories, which boosts Qwen3-VL-8B by 60% on VSI-Bench.

cs.CV↗

Geometry of Holomorphic One-forms on Smooth Projective Varieties

In this article, we show that any morphism $f$ from a smooth projective variety $X$ to a simple abelian variety $A$ is smooth, if and only if there exists a holomorphic 1-form $ω$ on $A$ such that $f^*ω$ has no zero. As the key ingredient in the proof, we show any $\mathbb{Z}$-homology fibre bundle morphism is without blow-up in codimension 0 in the sense of Sabbah. Furthermore, we investigate the structure of the spaces of holomorphic 1-forms with zeros, and show that they are linear for large classes of varieties. Also, we construct a delicate example of a smooth projective subvariety of an abelian variety for which the holomorphic 1-forms with positive dimensional zero loci do not form a linear subset. Finally, we study algebraic surfaces admitting holomorphic 1-forms that have zeros and do not arise from cohomology jump loci.

math.AG↗