arXiv subjects

Yulong Zhang

Publications and source records attributed to Yulong Zhang.

At least 19 recordsLinked to original sources

arXiv2026

WAM-TTT: Steering World-Action Models by Watching Human Play at Test Time

Steering robot foundation models (RFMs) toward new task variants or user-preferred behaviors remains challenging, often requiring additional robot demonstrations, task-specific fine-tuning, or long-context conditioning. We present WAM-TTT, a test-time training framework for steering world action models from raw human videos. Rather than treating human videos as trajectories to imitate, WAM-TTT absorbs them into a lightweight adaptive memory inside a frozen WAM through self-supervised video prediction. To make this memory useful for control, we introduce a meta-training stage that aligns human demonstrations with robot behaviors using paired human-robot data and a key--value memory reconstruction objective. At test time, only unlabeled human videos are required to adapt the memory, while the pretrained WAM remains frozen. This enables efficient and reusable steering without robot actions, human-side annotations, or task-specific fine-tuning, while preserving the generalization ability of the foundation model. Extensive experiments show that WAM-TTT consistently outperforms in-context human-video conditioning baselines across diverse manipulation tasks and generalization settings.

cs.RO↗

arXiv2026

Closing the Feedback Loop: From Experience Extraction to Insight Governance in Verbal Reinforcement Learning

Training-free verbal reinforcement learning enables LLM agents to learn from world feedback -- objective signals such as dynamic task outcomes, market returns, or demand forecasts -- by extracting verbal rules from experience and injecting them as context, updating the agent's behavior without parameter changes. However, in non-stationary environments these agents face a retention-forgetting dilemma: retaining stale insights causes negative transfer, while discarding them causes catastrophic forgetting when conditions recur. We identify four requirements for navigating this dilemma -- outcome-driven evaluation, persistent structured evidence, non-monotonic knowledge lifecycle, and compositional governance -- and show that existing methods invest heavily in experience extraction while underinvesting in insight governance. We propose a three-layer architecture -- rules, evidence, and skills -- connected by a feedback-driven curation loop that closes the governance gap. Rules capture distilled experience from world outcomes; evidence logs track each rule's reliability across episodes; skills govern which rules to apply, how to resolve conflicts, and when to abstain. On financial forecasting as a case study, where world feedback is naturally abundant, noisy, and non-stationary, we show that the same accumulated experience either degrades performance below the zero-shot baseline or dramatically improves accuracy and risk-adjusted returns, depending on whether the curation loop is present.

cs.AI↗

arXiv2026

Consensus Entropy: Harnessing Multi-VLM Agreement for Self-Verifying and Self-Improving OCR

Optical Character Recognition (OCR) is fundamental to Vision-Language Models (VLMs) and high-quality data generation for LLM training. Yet, despite progress in average OCR accuracy, state-of-the-art VLMs still struggle with detecting sample-level errors and lack effective unsupervised quality control. We introduce Consensus Entropy (CE), a training-free, model-agnostic metric that estimates output reliability by measuring inter-model agreement entropy. The core insight is that correct predictions converge in output space, while errors diverge. Based on CE, we develop CE-OCR, a lightweight multi-model framework that verifies outputs by ensemble agreement, selects the best outputs, and further improves efficiency through adaptive routing. Experiments demonstrate that CE is robust for quality verification, improving F1 scores by 42.1% over VLM-as-Judge. CE-OCR achieves consistent OCR gains, outperforming self-consistency and single-model baselines at the same cost. Notably, CE requires no training or supervision, enabling plug-and-play integration. Code: https://github.com/Aslan-yulong/consensus-entropy.

cs.CV↗

arXiv2026

Slack More, Predict Better: Proximal Relaxation for Probabilistic Latent Variable Model-based Soft Sensors

Nonlinear Probabilistic Latent Variable Models (NPLVMs) are a cornerstone of soft sensor modeling due to their capacity for uncertainty delineation. However, conventional NPLVMs are trained using amortized variational inference, where neural networks parameterize the variational posterior. While facilitating model implementation, this parameterization converts the distributional optimization problem within an infinite-dimensional function space to parameter optimization within a finite-dimensional parameter space, which introduces an approximation error gap, thereby degrading soft sensor modeling accuracy. To alleviate this issue, we introduce KProxNPLVM, a novel NPLVM that pivots to relaxing the objective itself and improving the NPLVM's performance. Specifically, we first prove the approximation error induced by the conventional approach. Based on this, we design the Wasserstein distance as the proximal operator to relax the learning objective, yielding a new variational inference strategy derived from solving this relaxed optimization problem. Based on this foundation, we provide a rigorous derivation of KProxNPLVM's optimization implementation, prove the convergence of our algorithm can finally sidestep the approximation error, and propose the KProxNPLVM by summarizing the abovementioned content. Finally, extensive experiments on synthetic and real-world industrial datasets are conducted to demonstrate the efficacy of the proposed KProxNPLVM.

cs.LG↗

arXiv2026

DexterCap: An Affordable and Automated System for Capturing Dexterous Hand-Object Manipulation

Capturing fine-grained hand-object interactions is challenging due to severe self-occlusion from closely spaced fingers and the subtlety of in-hand manipulation motions. Existing optical motion capture systems rely on expensive camera setups and extensive manual post-processing, while low-cost vision-based methods often suffer from reduced accuracy and reliability under occlusion. To address these challenges, we present DexterCap, a low-cost optical capture system for dexterous in-hand manipulation. DexterCap uses dense, character-coded marker patches to achieve robust tracking under severe self-occlusion, together with an automated reconstruction pipeline that requires minimal manual effort. With DexterCap, we introduce DexterHand, a dataset of fine-grained hand-object interactions covering diverse manipulation behaviors and objects, from simple primitives to complex articulated objects such as a Rubik's Cube. We release the dataset and code to support future research on dexterous hand-object interaction. Project website: https://pku-mocca.github.io/Dextercap-Page/

cs.GR↗

arXiv2026

COMETS: Coordinated Multi-Destination Video Transmission with In-Network Rate Adaptation

Large-scale video streaming events attract millions of simultaneous viewers, stressing existing delivery infrastructures. Client-driven adaptation reacts slowly to shared congestion, while server-based coordination introduces scalability bottlenecks and single points of failure. We present COMETS, a coordinated multi-destination video transmission framework that leverages information-centric networking principles such as request aggregation and in-network state awareness to enable scalable, fair, and adaptive rate control. COMETS introduces a novel range-interest protocol and distributed in-network decision process that aligns video quality across receiver groups while minimizing redundant transmissions. To achieve this, we develop a lightweight distributed optimization framework that guides per-hop quality adaptation without centralized control. Extensive emulation shows that COMETS consistently improves bandwidth utilization, fairness, and user-perceived quality of experience over DASH, MoQ, and ICN baselines, particularly under high concurrency. The results highlight COMETS as a practical, deployable approach for next-generation scalable video delivery.

cs.NI↗

arXiv2025

MIRNet: Integrating Constrained Graph-Based Reasoning with Pre-training for Diagnostic Medical Imaging

Automated interpretation of medical images demands robust modeling of complex visual-semantic relationships while addressing annotation scarcity, label imbalance, and clinical plausibility constraints. We introduce MIRNet (Medical Image Reasoner Network), a novel framework that integrates self-supervised pre-training with constrained graph-based reasoning. Tongue image diagnosis is a particularly challenging domain that requires fine-grained visual and semantic understanding. Our approach leverages self-supervised masked autoencoder (MAE) to learn transferable visual representations from unlabeled data; employs graph attention networks (GAT) to model label correlations through expert-defined structured graphs; enforces clinical priors via constraint-aware optimization using KL divergence and regularization losses; and mitigates imbalance using asymmetric loss (ASL) and boosting ensembles. To address annotation scarcity, we also introduce TongueAtlas-4K, a comprehensive expert-curated benchmark comprising 4,000 images annotated with 22 diagnostic labels--representing the largest public dataset in tongue analysis. Validation shows our method achieves state-of-the-art performance. While optimized for tongue diagnosis, the framework readily generalizes to broader diagnostic medical imaging tasks.

cs.CV↗

arXiv2025

OCR-Quality: A Human-Annotated Dataset for OCR Quality Assessment

We present OCR-Quality, a comprehensive human-annotated dataset designed for evaluating and developing OCR quality assessment methods. The dataset consists of 1,000 PDF pages converted to PNG images at 300 DPI, sampled from diverse real-world scenarios, including academic papers, textbooks, e-books, and multilingual documents. Each document has been processed using state-of-the-art Vision-Language Models (VLMs) and manually annotated with quality scores using a 4-level scoring system (1: Excellent, 2: Good, 3: Fair, 4: Poor). The dataset includes detailed source information, annotation guidelines, and representative cases across various difficulty levels. OCR-Quality addresses the critical need for reliable OCR quality assessment in real-world applications and provides a valuable benchmark for training and evaluating OCR verification systems. The dataset is publicly available at https://huggingface.co/datasets/Aslan-mingye/OCR-Quality .

cs.CV↗

arXiv2025

A Dual-Agent Adversarial Framework for Robust Generalization in Deep Reinforcement Learning

Recently, empowered with the powerful capabilities of neural networks, reinforcement learning (RL) has successfully tackled numerous challenging tasks. However, while these models demonstrate enhanced decision-making abilities, they are increasingly prone to overfitting. For instance, a trained RL model often fails to generalize to even minor variations of the same task, such as a change in background color or other minor semantic differences. To address this issue, we propose a dual-agent adversarial policy learning framework, which allows agents to spontaneously learn the underlying semantics without introducing any human prior knowledge. Specifically, our framework involves a game process between two agents: each agent seeks to maximize the impact of perturbing on the opponent's policy by producing representation differences for the same state, while maintaining its own stability against such perturbations. This interaction encourages agents to learn generalizable policies, capable of handling irrelevant features from the high-dimensional observations. Extensive experimental results on the Procgen benchmark demonstrate that the adversarial process significantly improves the generalization performance of both agents, while also being applied to various RL algorithms, e.g., Proximal Policy Optimization (PPO). With the adversarial framework, the RL agent outperforms the baseline methods by a significant margin, especially in hard-level tasks, marking a significant step forward in the generalization capabilities of deep reinforcement learning.

cs.LG↗

arXiv2025

NCV: A Node-Wise Consistency Verification Approach for Low-Cost Structured Error Localization in LLM Reasoning

Verifying multi-step reasoning in large language models is difficult due to imprecise error localization and high token costs. Existing methods either assess entire reasoning chains, suffering attention dilution, or rely on expensive multi-sampling. We introduce Node-wise Consistency Verification (NCV), a training-free framework that recasts verification as lightweight binary consistency checks at the node level. By decomposing the chain of thought into interconnected verification nodes, NCV precisely localizes errors and avoids unnecessary long-form generation. Experiments demonstrate that our approach enhances interpretability and efficiency, presenting a scalable solution for reliable LLM reasoning verification. On public datasets, NCV achieves a 10\% to 25\% improvement in F1 scores over baselines while utilizing $6\times$~$58\times$ fewer tokens than traditional methods like CoT-based verifiers.

cs.AI↗

arXiv2025

Come Together, But Not Right Now: A Progressive Strategy to Boost Low-Rank Adaptation

Low-rank adaptation (LoRA) has emerged as a leading parameter-efficient fine-tuning technique for adapting large foundation models, yet it often locks adapters into suboptimal minima near their initialization. This hampers model generalization and limits downstream operators such as adapter merging and pruning. Here, we propose CoTo, a progressive training strategy that gradually increases adapters' activation probability over the course of fine-tuning. By stochastically deactivating adapters, CoTo encourages more balanced optimization and broader exploration of the loss landscape. We provide a theoretical analysis showing that CoTo promotes layer-wise dropout stability and linear mode connectivity, and we adopt a cooperative-game approach to quantify each adapter's marginal contribution. Extensive experiments demonstrate that CoTo consistently boosts single-task performance, enhances multi-task merging accuracy, improves pruning robustness, and reduces training overhead, all while remaining compatible with diverse LoRA variants. Code is available at https://github.com/zwebzone/coto.

cs.LG↗

arXiv2025

ViFusion: In-Network Tensor Fusion for Scalable Video Feature Indexing

Large-scale video feature indexing in datacenters is critically dependent on efficient data transfer. Although in-network computation has emerged as a compelling strategy for accelerating feature extraction and reducing overhead in distributed multimedia systems, harnessing advanced networking resources at both the switch and host levels remains a formidable challenge. These difficulties are compounded by heterogeneous hardware, diverse application requirements, and complex multipath topologies. Existing methods focus primarily on optimizing inference for large neural network models using specialized collective communication libraries, which often face performance degradation in network congestion scenarios. To overcome these limitations, we present ViFusion, a communication aware tensor fusion framework that streamlines distributed video indexing by merging numerous small feature tensors into consolidated and more manageable units. By integrating an in-network computation module and a dedicated tensor fusion mechanism within datacenter environments, ViFusion substantially improves the efficiency of video feature indexing workflows. The deployment results show that ViFusion improves the throughput of the video retrieval system by 8--22 times with the same level of latency as state-of-the-art systems.

cs.MM↗

arXiv2025

Open Your Eyes: Vision Enhances Message Passing Neural Networks in Link Prediction

Message-passing graph neural networks (MPNNs) and structural features (SFs) are cornerstones for the link prediction task. However, as a common and intuitive mode of understanding, the potential of visual perception has been overlooked in the MPNN community. For the first time, we equip MPNNs with vision structural awareness by proposing an effective framework called Graph Vision Network (GVN), along with a more efficient variant (E-GVN). Extensive empirical results demonstrate that with the proposed frameworks, GVN consistently benefits from the vision enhancement across seven link prediction datasets, including challenging large-scale graphs. Such improvements are compatible with existing state-of-the-art (SOTA) methods and GVNs achieve new SOTA results, thereby underscoring a promising novel direction for link prediction.

cs.CV↗

arXiv2024

CopRA: A Progressive LoRA Training Strategy

Low-Rank Adaptation (LoRA) is a parameter-efficient technique for rapidly fine-tuning foundation models. In standard LoRA training dynamics, models tend to quickly converge to a local optimum near the initialization. However, this local optimum may not be ideal for out-of-distribution data or tasks such as merging and pruning. In this work, we propose a novel progressive training strategy for LoRA with random layer dropping. This strategy also optimizes the Shapley value of LoRA parameters in each layer, treating each layer as a player in a cooperative game. We refer to this method as Cooperative LoRA (CopRA). Our experimental results demonstrate that parameters trained with CopRA exhibit linear mode connectivity, which enables efficient model merging. This also paves the way for federated learning and multi-task learning via LoRA merging. Additionally, by optimizing the Shapley value, CopRA shows superior performance in pruning tasks.

cs.LG↗

arXiv2024

Secure Web Objects: Building Blocks for Metaverse Interoperability and Decentralization

This position paper explores how to support the Web's evolution through an underlying data-centric approach that better matches the data-orientedness of modern and emerging applications. We revisit the original vision of the Web as a hypermedia system that supports document composability and application interoperability via name-based data access. We propose the use of secure web objects (SWO), a data-oriented communication approach that can reduce complexity, centrality, and inefficiency, particularly for collaborative and local-first applications, such as the Metaverse and other collaborative applications. SWO are named, signed, application-defined objects that are secured independently of their containers or communications channels, an approach that leverages the results from over a decade-long data-centric networking research. This approach does not require intermediation by aggregators of identity, storage, and other services that are common today. We present a brief design overview, illustrated through prototypes for two editors of shared hypermedia documents: one for 3D and one for LaTeX. We also discuss our findings and suggest a roadmap for future research.

cs.NI↗

arXiv2024

Rethinking Guidance Information to Utilize Unlabeled Samples:A Label Encoding Perspective

Empirical Risk Minimization (ERM) is fragile in scenarios with insufficient labeled samples. A vanilla extension of ERM to unlabeled samples is Entropy Minimization (EntMin), which employs the soft-labels of unlabeled samples to guide their learning. However, EntMin emphasizes prediction discriminability while neglecting prediction diversity. To alleviate this issue, in this paper, we rethink the guidance information to utilize unlabeled samples. By analyzing the learning objective of ERM, we find that the guidance information for labeled samples in a specific category is the corresponding label encoding. Inspired by this finding, we propose a Label-Encoding Risk Minimization (LERM). It first estimates the label encodings through prediction means of unlabeled samples and then aligns them with their corresponding ground-truth label encodings. As a result, the LERM ensures both prediction discriminability and diversity, and it can be integrated into existing methods as a plugin. Theoretically, we analyze the relationships between LERM and ERM as well as EntMin. Empirically, we verify the superiority of the LERM under several label insufficient scenarios. The codes are available at https://github.com/zhangyl660/LERM.

cs.CV↗

arXiv2024

VLLaVO: Mitigating Visual Gap through LLMs

Recent advances achieved by deep learning models rely on the independent and identically distributed assumption, hindering their applications in real-world scenarios with domain shifts. To tackle this issue, cross-domain learning aims at extracting domain-invariant knowledge to reduce the domain shift between training and testing data. However, in visual cross-domain learning, traditional methods concentrate solely on the image modality, disregarding the potential benefits of incorporating the text modality. In this work, we propose VLLaVO, combining Vision language models and Large Language models as Visual cross-dOmain learners. VLLaVO uses vision-language models to convert images into detailed textual descriptions. A large language model is then finetuned on textual descriptions of the source/target domain generated by a designed instruction template. Extensive experimental results under domain generalization and unsupervised domain adaptation settings demonstrate the effectiveness of the proposed method.

cs.CV↗

arXiv2023

Domain-Guided Conditional Diffusion Model for Unsupervised Domain Adaptation

Limited transferability hinders the performance of deep learning models when applied to new application scenarios. Recently, Unsupervised Domain Adaptation (UDA) has achieved significant progress in addressing this issue via learning domain-invariant features. However, the performance of existing UDA methods is constrained by the large domain shift and limited target domain data. To alleviate these issues, we propose DomAin-guided Conditional Diffusion Model (DACDM) to generate high-fidelity and diversity samples for the target domain. In the proposed DACDM, by introducing class information, the labels of generated samples can be controlled, and a domain classifier is further introduced in DACDM to guide the generated samples for the target domain. The generated samples help existing UDA methods transfer from the source domain to the target domain more easily, thus improving the transfer performance. Extensive experiments on various benchmarks demonstrate that DACDM brings a large improvement to the performance of existing UDA methods.

cs.LG↗