arXiv subjects

Haoyu Zhang

Publications and source records attributed to Haoyu Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2025

Exo2Ego: Exocentric Knowledge Guided MLLM for Egocentric Video Understanding

AI personal assistants, deployed through robots or wearables, require embodied understanding to collaborate effectively with humans. However, current Multimodal Large Language Models (MLLMs) primarily focus on third-person (exocentric) vision, overlooking the unique challenges of first-person (egocentric) videos. Additionally, high acquisition costs limit data size, impairing MLLM performance. To address these challenges, we propose learning the mapping between exocentric and egocentric domains, leveraging the extensive exocentric knowledge within existing MLLMs to enhance egocentric video understanding. To this end, we introduce Ego-ExoClip, a pre-training dataset comprising 1.1M synchronized ego-exo clip-text pairs derived from Ego-Exo4D, together with the instruction-tuning dataset EgoIT, which is collected from multiple sources to enhance the model's instruction-following capabilities. Building upon the datasets, we propose a migration strategy and further design a progressive mapping learning pipeline with three stages: Demonstrator Self-Preparation, Demonstrator-Learner Guidance, and Learner Self-Practice. Extensive experiments across diverse egocentric tasks reveal that existing MLLMs perform inadequately in egocentric video understanding, while our model significantly outperforms these leading models.

cs.CV↗

arXiv2025

Towards Modality- and Sampling-Universal Learning Strategies for Accelerating Cardiovascular Imaging: Summary of the CMRxRecon2024 Challenge

Cardiovascular health is vital to human well-being, and cardiac magnetic resonance (CMR) imaging is considered the {clinical reference standard} for diagnosing cardiovascular disease. However, its adoption is hindered by long scan times, complex contrasts, and inconsistent quality. While deep learning methods perform well on specific CMR imaging {sequences}, they often fail to generalize across modalities and sampling schemes. The lack of benchmarks for high-quality, fast CMR image reconstruction further limits technology comparison and adoption. The CMRxRecon2024 challenge, attracting over 200 teams from 18 countries, addressed these issues with two tasks: generalization to unseen {modalities} and robustness to diverse undersampling patterns. We introduced the largest public multi-{modality} CMR raw dataset, an open benchmarking platform, and shared code. Analysis of the best-performing solutions revealed that prompt-based adaptation and enhanced physics-driven consistency enabled strong cross-scenario performance. These findings establish principles for generalizable reconstruction models and advance clinically translatable AI in cardiovascular imaging.

eess.IV↗

arXiv2025

Fortuity and R-charge concentration in the D1-D5 CFT

We investigate finite-$N$ BPS cohomology in the D1--D5 CFT, focusing on the sector of fortuitous classes. Analyzing the supercharge cochain complexes in the $N=2$ and $N=3$ theories, we construct several explicit fortuitous classes. We study the decomposition of these cohomology classes into ${\rm SU}(2)_a\times {\rm SU}(2)_b$ representations and conjecture that, at fixed holomorphic weight, those transforming in the largest representation are necessarily fortuitous. Our results also provide strong evidence that the $R$-charge concentration phenomenon extends to the D1--D5 CFT.

hep-th↗

arXiv2025

Task-Aware Retrieval Augmentation for Dynamic Recommendation

Dynamic recommendation systems aim to provide personalized suggestions by modeling temporal user-item interactions across time-series behavioral data. Recent studies have leveraged pre-trained dynamic graph neural networks (GNNs) to learn user-item representations over temporal snapshot graphs. However, fine-tuning GNNs on these graphs often results in generalization issues due to temporal discrepancies between pre-training and fine-tuning stages, limiting the model's ability to capture evolving user preferences. To address this, we propose TarDGR, a task-aware retrieval-augmented framework designed to enhance generalization capability by incorporating task-aware model and retrieval-augmentation. Specifically, TarDGR introduces a Task-Aware Evaluation Mechanism to identify semantically relevant historical subgraphs, enabling the construction of task-specific datasets without manual labeling. It also presents a Graph Transformer-based Task-Aware Model that integrates semantic and structural encodings to assess subgraph relevance. During inference, TarDGR retrieves and fuses task-aware subgraphs with the query subgraph, enriching its representation and mitigating temporal generalization issues. Experiments on multiple large-scale dynamic graph datasets demonstrate that TarDGR consistently outperforms state-of-the-art methods, with extensive empirical evidence underscoring its superior accuracy and generalization capabilities.

cs.IR↗

arXiv2025

Intention-Guided Cognitive Reasoning for Egocentric Long-Term Action Anticipation

Long-term action anticipation from egocentric video is critical for applications such as human-computer interaction and assistive technologies, where anticipating user intent enables proactive and context-aware AI assistance. However, existing approaches suffer from three key limitations: 1) underutilization of fine-grained visual cues from hand-object interactions, 2) neglect of semantic dependencies between verbs and nouns, and 3) lack of explicit cognitive reasoning, limiting generalization and long-term forecasting ability. To overcome these challenges, we propose INSIGHT, a unified two-stage framework for egocentric action anticipation. In the first stage, INSIGHT focuses on extracting semantically rich features from hand-object interaction regions and enhances action representations using a verb-noun co-occurrence matrix. In the second stage, it introduces a reinforcement learning-based module that simulates explicit cognitive reasoning through a structured process: visual perception (think) -> intention inference (reason) -> action anticipation (answer). Extensive experiments on Ego4D, EPIC-Kitchens-55, and EGTEA Gaze+ benchmarks show that INSIGHT achieves state-of-the-art performance, demonstrating its effectiveness and strong generalization capability.

cs.CV↗

arXiv2025

Robust High-Resolution Multi-Organ Diffusion MRI Using Synthetic-Data-Tuned Prompt Learning

Clinical adoption of multi-shot diffusion-weighted magnetic resonance imaging (multi-shot DWI) for body-wide tumor diagnostics is limited by severe motion-induced phase artifacts from respiration, peristalsis, and so on, compounded by multi-organ, multi-slice, multi-direction and multi-b-value complexities. Here, we introduce a reconstruction framework, LoSP-Prompt, that overcomes these challenges through physics-informed modeling and synthetic-data-driven prompt learning. We model inter-shot phase variations as a high-order Locally Smooth Phase (LoSP), integrated into a low-rank Hankel matrix reconstruction. Crucially, the algorithm's rank parameter is automatically set via prompt learning trained exclusively on synthetic abdominal DWI data emulating physiological motion. Validated across 10,000+ clinical images (43 subjects, 4 scanner models, 5 centers), LoSP-Prompt: (1) Achieved twice the spatial resolution of clinical single-shot DWI, enhancing liver lesion conspicuity; (2) Generalized to seven diverse anatomical regions (liver, kidney, sacroiliac, pelvis, knee, spinal cord, brain) with a single model; (3) Outperformed state-of-the-art methods in image quality, artifact suppression, and noise reduction (11 radiologists' evaluations on a 5-point scale, $p<0.05$), achieving 4-5 points (excellent) on kidney DWI, 4 points (good to excellent) on liver, sacroiliac and spinal cord DWI, and 3-4 points (good) on knee and tumor brain. The approach eliminates navigator signals and realistic data supervision, providing an interpretable, robust solution for high-resolution multi-organ multi-shot DWI. Its scanner-agnostic performance signifies transformative potential for precision oncology.

cs.CV↗

arXiv2025

Rethinking Graph Domain Adaptation: A Spectral Contrastive Perspective

Graph neural networks (GNNs) have achieved remarkable success in various domains, yet they often struggle with domain adaptation due to significant structural distribution shifts and insufficient exploration of transferable patterns. One of the main reasons behind this is that traditional approaches do not treat global and local patterns discriminatingly so that some local details in the graph may be violated after multi-layer GNN. Our key insight is that domain shifts can be better understood through spectral analysis, where low-frequency components often encode domain-invariant global patterns, and high-frequency components capture domain-specific local details. As such, we propose FracNet (\underline{\textbf{Fr}}equency \underline{\textbf{A}}ware \underline{\textbf{C}}ontrastive Graph \underline{\textbf{Net}}work) with two synergic modules to decompose the original graph into high-frequency and low-frequency components and perform frequency-aware domain adaption. Moreover, the blurring boundary problem of domain adaptation is improved by integrating with a contrastive learning framework. Besides the practical implication, we also provide rigorous theoretical proof to demonstrate the superiority of FracNet. Extensive experiments further demonstrate significant improvements over state-of-the-art approaches.

cs.LG↗

arXiv2025

Optimal-PhiBE: A PDE-based Model-free framework for Continuous-time Reinforcement Learning

This paper addresses continuous-time reinforcement learning (CTRL) where the system dynamics are governed by an unknown stochastic differential equation, and only discrete-time observations are available. Existing approaches face limitations: model-based PDE methods suffer from non-identifiability, while model-free methods based on the discrete-time optimal Bellman equation (Optimal-BE) suffer from large discretization errors that are highly sensitive to both the system dynamics and the reward structure. To overcome these challenges, we introduce Optimal-PhiBE, a formulation that integrates discrete-time information into a continuous-time PDE, combining the strength of both existing frameworks while mitigating their limitations. Optimal-PhiBE exhibits smaller discretization errors when the uncontrolled system evolves slowly, and demonstrates reduced sensitivity to oscillatory reward structures, and enables model-free algorithms that bypass explicit dynamics estimation. In the linear-quadratic regulator (LQR) setting, sharp error bounds are established for both Optimal-PhiBE and Optimal-BE. The results show that Optimal-PhiBE exactly recovers the optimal policy in the undiscounted case and substantially outperforms Optimal-BE when the problem is weakly discounted or control-dominant. Furthermore, we extend Optimal-PhiBE to higher orders, providing increasingly accurate approximations. A model-free policy iteration algorithm is proposed to solve the Optimal-PhiBE directly from trajectory data. Numerical experiments are conducted to verify the theoretical findings.

math.OC↗

arXiv2025

STRAP: Spatio-Temporal Pattern Retrieval for Out-of-Distribution Generalization

Spatio-Temporal Graph Neural Networks (STGNNs) have emerged as a powerful tool for modeling dynamic graph-structured data across diverse domains. However, they often fail to generalize in Spatio-Temporal Out-of-Distribution (STOOD) scenarios, where both temporal dynamics and spatial structures evolve beyond the training distribution. To address this problem, we propose an innovative Spatio-Temporal Retrieval-Augmented Pattern Learning framework,STRAP, which enhances model generalization by integrating retrieval-augmented learning into the STGNN continue learning pipeline. The core of STRAP is a compact and expressive pattern library that stores representative spatio-temporal patterns enriched with historical, structural, and semantic information, which is obtained and optimized during the training phase. During inference, STRAP retrieves relevant patterns from this library based on similarity to the current input and injects them into the model via a plug-and-play prompting mechanism. This not only strengthens spatio-temporal representations but also mitigates catastrophic forgetting. Moreover, STRAP introduces a knowledge-balancing objective to harmonize new information with retrieved knowledge. Extensive experiments across multiple real-world streaming graph datasets show that STRAP consistently outperforms state-of-the-art STGNN baselines on STOOD tasks, demonstrating its robustness, adaptability, and strong generalization capability without task-specific fine-tuning.

cs.LG↗

arXiv2025

Spatial Understanding from Videos: Structured Prompts Meet Simulation Data

Visual-spatial understanding, the ability to infer object relationships and layouts from visual input, is fundamental to downstream tasks such as robotic navigation and embodied interaction. However, existing methods face spatial uncertainty and data scarcity, limiting the 3D spatial reasoning capability of pre-trained vision-language models (VLMs). To address these challenges, we present a unified framework for enhancing 3D spatial reasoning in pre-trained VLMs without modifying their architecture. This framework combines SpatialMind, a structured prompting strategy that decomposes complex scenes and questions into interpretable reasoning steps, with ScanForgeQA, a scalable question-answering dataset built from diverse 3D simulation scenes through an automated construction process designed for fine-tuning. Extensive experiments across multiple benchmarks demonstrate the individual and combined effectiveness of our prompting and fine-tuning strategies, and yield insights that may inspire future research on visual-spatial understanding.

cs.CV↗

arXiv2025

Large VLM-based Stylized Sports Captioning

The advent of large (visual) language models (LLM / LVLM) have led to a deluge of automated human-like systems in several domains including social media content generation, search and recommendation, healthcare prognosis, AI assistants for cognitive tasks etc. Although these systems have been successfully integrated in production; very little focus has been placed on sports, particularly accurate identification and natural language description of the game play. Most existing LLM/LVLMs can explain generic sports activities, but lack sufficient domain-centric sports' jargon to create natural (human-like) descriptions. This work highlights the limitations of existing SoTA LLM/LVLMs for generating production-grade sports captions from images in a desired stylized format, and proposes a two-level fine-tuned LVLM pipeline to address that. The proposed pipeline yields an improvement > 8-10% in the F1, and > 2-10% in BERT score compared to alternative approaches. In addition, it has a small runtime memory footprint and fast execution time. During Super Bowl LIX the pipeline proved its practical application for live professional sports journalism; generating highly accurate and stylized captions at the rate of 6 images per 3-5 seconds for over 1000 images during the game play.

cs.CV↗

arXiv2025

Boosting Action-Information via a Variational Bottleneck on Unlabelled Robot Videos

Learning from demonstrations (LfD) typically relies on large amounts of action-labeled expert trajectories, which fundamentally constrains the scale of available training data. A promising alternative is to learn directly from unlabeled video demonstrations. However, we find that existing methods tend to encode latent actions that share little mutual information with the true robot actions, leading to suboptimal control performance. To address this limitation, we introduce a novel framework that explicitly maximizes the mutual information between latent actions and true actions, even in the absence of action labels. Our method leverage the variational information-bottleneck to extract action-relevant representations while discarding task-irrelevant information. We provide a theoretical analysis showing that our objective indeed maximizes the mutual information between latent and true actions. Finally, we validate our approach through extensive experiments: first in simulated robotic environments and then on real-world robotic platforms, the experimental results demonstrate that our method significantly enhances mutual information and consistently improves policy performance.

cs.RO↗

arXiv2025

RideAgent: An LLM-Enhanced Optimization Framework for Automated Taxi Fleet Operations

Efficient management of electric ride-hailing fleets, particularly pre-allocation and pricing during peak periods to balance spatio-temporal supply and demand, is crucial for urban traffic efficiency. However, practical challenges include unpredictable demand and translating diverse, qualitative managerial objectives from non-expert operators into tractable optimization models. This paper introduces RideAgent, an LLM-powered agent framework that automates and enhances electric ride-hailing fleet management. First, an LLM interprets natural language queries from fleet managers to formulate corresponding mathematical objective functions. These user-defined objectives are then optimized within a Mixed-Integer Programming (MIP) framework, subject to the constraint of maintaining high operational profit. The profit itself is a primary objective, estimated by an embedded Random Forest (RF) model leveraging exogenous features. To accelerate the solution of this MIP, a prompt-guided LLM analyzes a small sample of historical optimal decision data to guide a variable fixing strategy. Experiments on real-world data show that the LLM-generated objectives achieve an 86% text similarity to standard formulations in a zero-shot setting. Following this, the LLM-guided variable fixing strategy reduces computation time by 53.15% compared to solving the full MIP with only a 2.42% average optimality gap. Moreover, this variable fixing strategy outperforms five cutting plane methods by 42.3% time reduction with minimal compromise to solution quality. RideAgent offers a robust and adaptive automated framework for objective modeling and accelerated optimization. This framework empowers non-expert fleet managers to personalize operations and improve urban transportation system performance.

math.OC↗

arXiv2025

SCING:Towards More Efficient and Robust Person Re-Identification through Selective Cross-modal Prompt Tuning

Recent advancements in adapting vision-language pre-training models like CLIP for person re-identification (ReID) tasks often rely on complex adapter design or modality-specific tuning while neglecting cross-modal interaction, leading to high computational costs or suboptimal alignment. To address these limitations, we propose a simple yet effective framework named Selective Cross-modal Prompt Tuning (SCING) that enhances cross-modal alignment and robustness against real-world perturbations. Our method introduces two key innovations: Firstly, we proposed Selective Visual Prompt Fusion (SVIP), a lightweight module that dynamically injects discriminative visual features into text prompts via a cross-modal gating mechanism. Moreover, the proposed Perturbation-Driven Consistency Alignment (PDCA) is a dual-path training strategy that enforces invariant feature alignment under random image perturbations by regularizing consistency between original and augmented cross-modal embeddings. Extensive experiments are conducted on several popular benchmarks covering Market1501, DukeMTMC-ReID, Occluded-Duke, Occluded-REID, and P-DukeMTMC, which demonstrate the impressive performance of the proposed method. Notably, our framework eliminates heavy adapters while maintaining efficient inference, achieving an optimal trade-off between performance and computational overhead. The code will be released upon acceptance.

cs.CV↗

arXiv2025

Technical Report for Ego4D Long-Term Action Anticipation Challenge 2025

In this report, we present a novel three-stage framework developed for the Ego4D Long-Term Action Anticipation (LTA) task. Inspired by recent advances in foundation models, our method consists of three stages: feature extraction, action recognition, and long-term action anticipation. First, visual features are extracted using a high-performance visual encoder. The features are then fed into a Transformer to predict verbs and nouns, with a verb-noun co-occurrence matrix incorporated to enhance recognition accuracy. Finally, the predicted verb-noun pairs are formatted as textual prompts and input into a fine-tuned large language model (LLM) to anticipate future action sequences. Our framework achieves first place in this challenge at CVPR 2025, establishing a new state-of-the-art in long-term action prediction. Our code will be released at https://github.com/CorrineQiu/Ego4D-LTA-Challenge-2025.

cs.CV↗

arXiv2025

Designer polyradical nanographenes with strong spin entanglement and perturbation resilience via Clar's goblet extension

Polyradical nanographenes featuring strong spin entanglement and robust many-body spin states against external magnetic perturbations not only enable the exploration of correlated quantum magnetism at the molecular scale, but also constitute promising candidates for developing molecular qubits with chemical tunability and building scalable quantum networks. Here, we employed a predictive design strategy to achieve the on-surface synthesis of two homologues of Clar goblet, C62H22 and C76H26, via lateral and vertical extensions of the parent structure, respectively. Vertical extension increases the number of topologically frustrated zero-energy modes, which scale linearly with the total number of benzene ring rows. In contrast, the lateral extension enhances electron-electron interactions, leading to the emergence of additional radical states beyond those predicted by the topological zero-energy modes. Consequently, both structures exhibit correlated tetraradical character and a many-body singlet ground state as confirmed by multireference theoretical calculations. These magnetic states arise from unique magnetic origins and also display distinct resilience to external perturbations, which can be experimentally validated using nickelocene-functionalized scanning probe techniques. Our work presents a general strategy for rational design of highly entangled polyradical nanographenes with tunable spin numbers and resilience of their many-body spin states to perturbations, opening exciting possibilities for exploring novel correlated spin phases in molecular systems and advancing quantum information technologies.

cond-mat.mes-hall↗

arXiv2025

OSGNet @ Ego4D Episodic Memory Challenge 2025

In this report, we present our champion solutions for the three egocentric video localization tracks of the Ego4D Episodic Memory Challenge at CVPR 2025. All tracks require precise localization of the interval within an untrimmed egocentric video. Previous unified video localization approaches often rely on late fusion strategies, which tend to yield suboptimal results. To address this, we adopt an early fusion-based video localization model to tackle all three tasks, aiming to enhance localization accuracy. Ultimately, our method achieved first place in the Natural Language Queries, Goal Step, and Moment Queries tracks, demonstrating its effectiveness. Our code can be found at https://github.com/Yisen-Feng/OSGNet.

cs.CV↗

arXiv2025

HCQA-1.5 @ Ego4D EgoSchema Challenge 2025

In this report, we present the method that achieves third place for Ego4D EgoSchema Challenge in CVPR 2025. To improve the reliability of answer prediction in egocentric video question answering, we propose an effective extension to the previously proposed HCQA framework. Our approach introduces a multi-source aggregation strategy to generate diverse predictions, followed by a confidence-based filtering mechanism that selects high-confidence answers directly. For low-confidence cases, we incorporate a fine-grained reasoning module that performs additional visual and contextual analysis to refine the predictions. Evaluated on the EgoSchema blind test set, our method achieves 77% accuracy on over 5,000 human-curated multiple-choice questions, outperforming last year's winning solution and the majority of participating teams. Our code will be added at https://github.com/Hyu-Zhang/HCQA.

cs.CV↗