arXiv subjects

Weiming Li

Publications and source records attributed to Weiming Li.

At least 19 recordsLinked to original sources

arXiv2026

Candidate Generation and Definition-Guided Verification for Sentence-Level Depression Symptom Recognition

Sentence-level recognition of depression symptoms is challenging because similar expressions can differ in symptom relevance, and language-model inference is insufficiently grounded in diagnostic definitions. This study proposes a two-stage framework separating symptom-candidate generation from definition-grounded verification. A contrastively fine-tuned sentence encoder generates a symptom candidate per sentence, and a fine-tuned language model verifies whether the candidate is present or absent using the sentence, its context, and a candidate-specific diagnostic definition, checking its judgment against that definition before answering. Evaluated against encoder, inference-based, medical, and general LLM baselines and a matched single-stage supervised classifier, the proposed pipeline attains the best accuracy and F1 scores of all methods, with rationales matching expert-authored annotations. A preliminary clinical audit indicates moderate alignment with diagnostic definitions, with explanation quality strongly dependent on prediction correctness. The results support decomposing symptom recognition into candidate generation and definition-grounded verification, though performance remains limited for rare categories.

cs.CL↗

arXiv2026

LocalLSTC: A Long Short-Term Control Architecture for Locally Deployed GUI Agents

Modern GUI-agent frameworks achieve strong desktop task performance with frontier API models, yet persistent control information often remains implicit in growing interaction trajectories. At each step, the planner reconstructs the active task stage, accumulated evidence, and runtime feedback before deciding the next action. This dependence becomes more pronounced under weaker local reasoning backbones. Across four representative state-of-the-art frameworks, replacing GPT-5 with Qwen3.5-9B reduces average OSWorld SR-100 from 60.9\% to 37.7\%. Trajectory annotation further identifies at least one control failure in 91.6\% of failed trajectories. To address this problem, we introduce LocalLSTC, a training-free architecture that organizes control by temporal scope, maintaining persistent cross-step state to guide short-term execution commitments. Long-Term Control maintains the active subgoal, subgoal-aligned evidence, and runtime feedback across interactions, while Short-Term Execution realizes bounded commitments for the current step. Long-to-Short Planning forms each commitment from persistent state, and Short-to-Long Control integrates execution outcomes back into that state for progress assessment, recovery, and termination. With Qwen3.6-27B, LocalLSTC reaches 64.7\% SR-100 on OSWorld and 65.3\% on WindowsAgentArena, outperforming the strongest prior local results on both benchmarks. Ablations further support contributions from mechanisms on both sides of execution. These findings identify temporal organization of control information as a distinct architectural dimension for locally deployed GUI agents.

cs.AI↗

arXiv2026

R4DSG: Relative 4D Scene Graph Memory for Object-Centric Question Answering in Long Egocentric Video

Long-horizon egocentric video is a rich substrate for wearable AI assistants, but object-centric questions such as where an item was moved, when it last changed state, or why it was relocated remain difficult because caption- and transcript-based memories rarely preserve persistent object identity or structured spatial change. Existing long-video QA methods mainly emphasize temporal grounding and clip retrieval, while prior 3D scene-graph methods typically assume stronger geometry than free-motion wearable RGB video provides, including point clouds, RGB-D input, posed views, sparse reconstruction, or reconstructed scenes. R4DSG introduces a relative 4D scene graph memory for long egocentric video. Instead of storing raw graph sequences, R4DSG converts video into compact queryable memory entries indexed by time, place, persistent objects, anchor-relative change, and local interaction context. The main idea is to separate stable anchors from dynamic objects, maintain persistent object identity across frames, and represent object state through anchor-relative transitions rather than a globally aligned world model. Built on recent RGB-only advances in promptable video segmentation, temporal propagation, and relative 3D lifting, the method produces a retrieval-ready memory directly usable for long-horizon question answering. Evaluation on a 255-question object-related subset from EgoLifeQA shows, under question-only retrieval, a 6.7-point overall gain over EgoRAG-Text and a 12.5-point gain on when questions, which highlights the value of temporally organized object memory. These results position relative 4D scene graphs as a practical memory substrate for wearable assistants, AR systems, and embodied multimedia agents. GitHub Page: https://dualtransparency.github.io/R4DSG/.

cs.CV↗

arXiv2026

From Centrality Discounts to Centrality Premia: Interoperability and Platform Competition in Social Networks

We study how interoperability reshapes competitive price discrimination when consumers are embedded in a social network. Two differentiated platforms set personalized prices; consumers benefit from neighbors' consumption of the same platform and, under interoperability, of the rival. Equilibrium prices obtain in closed form for arbitrary networks and contain a network-position term, proportional to Katz-Bonacich centrality, whose sign is determined by whether interoperability exceeds product substitutability. Below this threshold, platforms contest central consumers and grant centrality discounts; above it, central consumers become gateways to a shared cross-platform network and pay premia; at the threshold, prices are independent of network position. Interoperability softens price competition, can make platforms favor denser consumer networks, and reverses which side of the market gains from price discrimination.

econ.TH↗

arXiv2026

Co-VLA: Coordination-Aware Structured Action Modeling for Dual-Arm Vision-Language-Action Systems

Vision-language-action (VLA) models show strong capabilities in single and dual-arm robotic manipulation. Prior works show coordinated bimanual behaviors can emerge from end-to-end learning, leveraging large vision-language backbones with continuous action prediction. However, as bimanual tasks become tightly coupled and execution constraints become critical, implicit coordination alone is insufficient to ensure reliable, interpretable, and stable behavior. In this work, we propose Co-VLA, a coordination-aware bimanual manipulation framework introducing explicit structural priors into VLA models. We instantiate our method on a state-of-the-art vision-language backbone by replacing its monolithic action head with a Structured Action Expert (SAE) designed for bimanual coordination. Specifically, we introduce explicit structure at the action generation level with a modular coordination-aware loss that shapes shared and residual latents according to task-specific structures. The shared latent encodes task-level coordination intent, while residual latents capture execution adjustments for each arm. At deployment, a Latent-Aware Controller (LAC) interprets the learned representations to modulate synchronization strength, execution asymmetry, smoothness, and safety constraints in real time. LAC operates at the joint-command level and remains compatible with standard control pipelines without requiring force or impedance control. Experiments across simulation and real-world benchmarks show Co-VLA significantly outperforms monolithic baselines, achieving a 27% success rate gain in tight-coordination tasks, more than doubling performance in OOD real-world scenarios (from 13% to 27%), and reducing task completion time by up to 25%.

cs.RO↗

arXiv2026

Mem-World: Memory-Augmented Action-Conditioned World Models for Persistent Robot Manipulation

Action-conditioned world models have emerged as a promising paradigm for robot learning, offering a scalable alternative to costly real-world experimentation by generating action-consistent video rollouts. However, persistent world modeling remains challenging in manipulation: frequent end-effector occlusions and rapid wrist-camera motion make the current observation insufficient for predicting future views, causing models to forget or hallucinate scene details seen in earlier frames. Existing memory retrieval strategies often fail to identify informative history in dynamic manipulation scenarios. To address this limitation, we propose Mem-World, a memory-augmented multi-view action-conditioned world model. At its core, we present W-VMem, a 4D wrist-view-centered surfel-indexed memory that anchors historical observations to temporally evolving surface elements. By explicitly modeling when and where scene elements are observed, W-VMem enables geometry-aware retrieval of relevant history frames conditioned on future actions. During generation, relevant history frames are selected via surfel-based rendering and scoring, providing informative and non-redundant context for prediction. Extensive experiments show that Mem-World generates persistent rollouts in complex manipulation scenarios, enables more reliable policy evaluation than Ctrl-World, improving the Pearson correlation with real-world performance by 14.5\%, and supports effective policy improvement through synthetic data generation, increasing success rates from 58\% to 72\% on long-horizon tasks.

cs.CV↗

arXiv2026

DAM-VLA: A Dynamic Action Model-Based Vision-Language-Action Framework for Robot Manipulation

In dynamic environments such as warehouses, hospitals, and homes, robots must seamlessly transition between gross motion and precise manipulations to complete complex tasks. However, current Vision-Language-Action (VLA) frameworks, largely adapted from pre-trained Vision-Language Models (VLMs), often struggle to reconcile general task adaptability with the specialized precision required for intricate manipulation. To address this challenge, we propose DAM-VLA, a dynamic action model-based VLA framework. DAM-VLA integrates VLM reasoning with diffusion-based action models specialized for arm and gripper control. Specifically, it introduces (i) an action routing mechanism, using task-specific visual and linguistic cues to select appropriate action models (e.g., arm movement or gripper manipulation), (ii) a dynamic action model that fuses high-level VLM cognition with low-level visual features to predict actions, and (iii) a dual-scale action weighting mechanism that enables dynamic coordination between the arm-movement and gripper-manipulation models. Across extensive evaluations, DAM-VLA achieves superior success rates compared to state-of-the-art VLA methods in simulated (SIMPLER, FurnitureBench) and real-world settings, showing robust generalization from standard pick-and-place to demanding long-horizon and contact-rich tasks.

cs.RO↗

arXiv2026

High-Dimensional Precision Matrix Quadratic Forms: Estimation Framework for $p > n$

We propose a novel estimation framework for quadratic functionals of precision matrices in high-dimensional settings, particularly in regimes where the feature dimension $p$ exceeds the sample size $n$. Traditional moment-based estimators with bias correction remain consistent when $p n$, highlighting a fundamental distinction between the two regimes due to rank deficiency and high-dimensional complexity. Our approach resolves these issues by combining a spectral-moment representation with constrained optimization, resulting in consistent estimation under mild moment conditions. The proposed framework provides a unified approach for inference on a broad class of high-dimensional statistical measures. We illustrate its utility through two representative examples: the optimal Sharpe ratio in portfolio optimization and the multiple correlation coefficient in regression analysis. Simulation studies demonstrate that the proposed estimator effectively overcomes the fundamental $p>n$ barrier where conventional methods fail.

stat.ME↗

arXiv2025

BF-APNN: A Low-Memory Method for Accelerating the Solution of Radiative Transfer Equations

The Radiative Transfer Equations (RTEs) exhibit high dimensionality and multiscale characteristics, rendering conventional numerical methods computationally intensive. Existing deep learning methods perform well in low-dimensional or linear RTEs, but still face many challenges with high-dimensional or nonlinear RTEs. To overcome these challenges, we propose the Basis Function Asymptotically Preserving Neural Network (BF-APNN), a framework that inherits the advantages of Radiative Transfer Asymptotically Preserving Neural Network (RT-APNN) and accelerates the solution process. By employing basis function expansion on the microscopic component, derived from micro-macro decomposition, BF-APNN effectively mitigates the computational burden associated with evaluating high-dimensional integrals during training. Numerical experiments, which involve challenging RTE scenarios featuring, nonlinearity, discontinuities, and multiscale behavior, demonstrate that BF-APNN substantially reduces training time compared to RT-APNN while preserving high solution accuracy. Moreover, BF-APNN exhibits superior performance in addressing complex, high-dimensional RTE problems, underscoring its potential as a robust tool for radiative transfer computations.

physics.comp-ph↗

arXiv2025

On Convergence Rates of Spiked Eigenvalue Estimates: A General Study of Global and Local Laws in Sample Covariance Matrices

This paper investigates global and local laws for sample covariance matrices with general growth rates of dimensions. The sample size $N$ and population dimension $M$ can have the same order in logarithm, which implies that their ratio $M/N$ can approach zero, a constant, or infinity. These theories are utilized to determine the convergence rate of spiked eigenvalue estimates.

math.ST↗

arXiv2025

Question-Adaptive Graph Learning for Multi-hop Retrieval Augmented Generation

Retrieval-augmented generation (RAG) has demonstrated its ability to enhance Large Language Models (LLMs) by integrating external knowledge sources. However, multi-hop questions, which require the identification of multiple knowledge targets to form a synthesized answer, raise new challenges for RAG systems. Under the multi-hop settings, existing methods often struggle to fully understand the questions with complex semantic structures and are susceptible to irrelevant noise during the retrieval of multiple information targets. To address these limitations, we propose a novel graph representation learning framework for multi-hop question retrieval. We first introduce a Multi-information Level Knowledge Graph (Multi-L KG) to model various information levels for a more comprehensive understanding of multi-hop questions. Based on this, we design a Question-Adaptive Graph Neural Network (Quest-GNN) for representation learning on the Multi-L KG. Quest-GNN employs intra/inter-level message passing mechanisms, and in each message passing the information aggregation is guided by the question, which not only facilitates multi-granular information aggregation but also significantly reduces the impact of noise. To enhance its ability to learn robust representations, we further propose two synthesized data generation strategies for pre-training the Quest-GNN. Extensive experimental results demonstrate the effectiveness of our framework in multi-hop scenarios, especially in high-hop questions the improvement can reach 33.8\%. The code is available at: https://github.com/Jerry2398/QSGNN.

cs.LG↗

arXiv2025

How Auxiliary Reasoning Unleashes GUI Grounding in VLMs

Graphical user interface (GUI) grounding is a fundamental task for building GUI agents. However, general vision-language models (VLMs) struggle with this task due to a lack of specific optimization. We identify a key gap in this paper: while VLMs exhibit significant latent grounding potential, as demonstrated by their performance measured by Pointing Game, they underperform when tasked with outputting explicit coordinates. To address this discrepancy and bypass the high data and annotation costs of current fine-tuning approaches, we propose three zero-shot auxiliary reasoning methods. By providing explicit spatial cues such as axes, grids and labeled intersections as part of the input image, these methods enable VLMs to better articulate their implicit spatial understanding capabilities. We evaluate these methods on four GUI grounding benchmarks across seven open-source and proprietary VLMs. Experimental results show substantial gains from auxiliary reasoning. Mark-Grid Scaffold boosts Gemini-3.1-Pro from 11.72\% under direct inference to 95.20\% on ScreenSpot-v2, achieves state-of-the-art performance on ScreenSpot, and approaches the strongest fine-tuned methods on ScreenSpot-v2 and UI-I2E-Bench. Our code is available at https://github.com/liweim/AuxiliaryReasoning.

cs.CV↗

arXiv2025

MoSE: Skill-by-Skill Mixture-of-Experts Learning for Embodied Autonomous Machines

To meet the growing demand for smarter, faster, and more efficient embodied AI solutions, we introduce a novel Mixture-of-Expert (MoE) method that significantly boosts reasoning and learning efficiency for embodied autonomous systems. General MoE models demand extensive training data and complex optimization, which limits their applicability in embodied AI such as autonomous driving (AD) and robotic manipulation. In this work, we propose a skill-oriented MoE called MoSE, which mimics the human learning and reasoning process skill-by-skill, step-by-step. We introduce a skill-oriented routing mechanism that begins with defining and annotating specific skills, enabling experts to identify the necessary competencies for various scenarios and reasoning tasks, thereby facilitating skill-by-skill learning. To better align with multi-step planning in human reasoning and in end-to-end driving models, we build a hierarchical skill dataset and pretrain the router to encourage the model to think step-by-step. Unlike other multi-round dialogues, MoSE integrates valuable auxiliary tasks (e.g. perception-prediction-planning for AD, and high-level and low-level planning for robots) in one single forward process without introducing any extra computational cost. With less than 3B sparsely activated parameters, our model effectively grows more diverse expertise and outperforms models on both AD corner-case reasoning tasks and robot reasoning tasks with less than 40% of the parameters.

cs.AI↗

arXiv2025

Debiased distributed PCA under high dimensional spiked model

We study distributed principal component analysis (PCA) in high-dimensional settings under the spiked model. In such regimes, sample eigenvectors can deviate significantly from population ones, introducing a persistent bias. Existing distributed PCA methods are sensitive to this bias, particularly when the number of machines is small. Their consistency typically relies on the number of machines tending to infinity. We propose a debiased distributed PCA algorithm that corrects the local bias before aggregation and incorporates a sparsity-detection step to adaptively handle sparse and non-sparse eigenvectors. Theoretically, we establish the consistency of our estimator under much weaker conditions compared to existing literature. In particular, our approach does not require symmetric innovations and only assumes a finite sixth moment. Furthermore, our method generally achieves smaller estimation error, especially when the number of machines is small. Empirically, extensive simulations and real data experiments demonstrate that our method consistently outperforms existing distributed PCA approaches. The advantage is especially prominent when the leading eigenvectors are sparse or the number of machines is limited. Our method and theoretical analysis are also applicable to the sample correlation matrix.

stat.ME↗

arXiv2025

DomainCQA: Crafting Knowledge-Intensive QA from Domain-Specific Charts

Chart Question Answering (CQA) evaluates Multimodal Large Language Models (MLLMs) on visual understanding and reasoning over chart data. However, existing benchmarks mostly test surface-level parsing, such as reading labels and legends, while overlooking deeper scientific reasoning. We propose DomainCQA, a framework for constructing domain-specific CQA benchmarks that emphasize both visual comprehension and knowledge-intensive reasoning. It integrates complexity-aware chart selection, multitier QA generation, and expert validation. Applied to astronomy, DomainCQA yields AstroChart, a benchmark of 1,690 QA pairs over 482 charts, exposing persistent weaknesses in fine-grained perception, numerical reasoning, and domain knowledge integration across 21 MLLMs. Fine-tuning on AstroChart improves performance across fundamental and advanced tasks. Pilot QA sets in biochemistry, economics, medicine, and social science further demonstrate DomainCQA's generality. Together, our results establish DomainCQA as a unified pipeline for constructing and augmenting domain-specific chart reasoning benchmarks.

cs.CL↗

arXiv2025

MapFusion: A Novel BEV Feature Fusion Network for Multi-modal Map Construction

Map construction task plays a vital role in providing precise and comprehensive static environmental information essential for autonomous driving systems. Primary sensors include cameras and LiDAR, with configurations varying between camera-only, LiDAR-only, or camera-LiDAR fusion, based on cost-performance considerations. While fusion-based methods typically perform best, existing approaches often neglect modality interaction and rely on simple fusion strategies, which suffer from the problems of misalignment and information loss. To address these issues, we propose MapFusion, a novel multi-modal Bird's-Eye View (BEV) feature fusion method for map construction. Specifically, to solve the semantic misalignment problem between camera and LiDAR BEV features, we introduce the Cross-modal Interaction Transform (CIT) module, enabling interaction between two BEV feature spaces and enhancing feature representation through a self-attention mechanism. Additionally, we propose an effective Dual Dynamic Fusion (DDF) module to adaptively select valuable information from different modalities, which can take full advantage of the inherent information between different modalities. Moreover, MapFusion is designed to be simple and plug-and-play, easily integrated into existing pipelines. We evaluate MapFusion on two map construction tasks, including High-definition (HD) map and BEV map segmentation, to show its versatility and effectiveness. Compared with the state-of-the-art methods, MapFusion achieves 3.6% and 6.2% absolute improvements on the HD map construction and BEV map segmentation tasks on the nuScenes dataset, respectively, demonstrating the superiority of our approach.

cs.CV↗

arXiv2024

An asymptotic-preserving IMEX PN method for the gray model of the radiative transfer equation

An asymptotic-preserving (AP) implicit-explicit PN numerical scheme is proposed for the gray model of the radiative transfer equation, where the first- and second-order numerical schemes are discussed for both the linear and nonlinear models. The AP property of this numerical scheme is proved theoretically and numerically, while the numerical stability of the linear model is verified by Fourier analysis. Several classical benchmark examples are studied to validate the efficiency of this numerical scheme.

math.NA↗

arXiv2024

Terahertz Channels in Atmospheric Conditions: Propagation Characteristics and Security Performance

With the growing demand for higher wireless data rates, the interest in extending the carrier frequency of wireless links to the terahertz (THz) range has significantly increased. For long-distance outdoor wireless communications, THz channels may suffer substantial power loss and security issues due to atmospheric weather effects. It is crucial to assess the impact of weather on high-capacity data transmission to evaluate wireless system link budgets and performance accurately. In this article, we provide an insight into the propagation characteristics of THz channels under atmospheric conditions and the security aspects of THz communication systems in future applications. We conduct a comprehensive survey of our recent research and experimental findings on THz channel transmission and physical layer security, synthesizing and categorizing the state-of-the-art research in this domain. Our analysis encompasses various atmospheric phenomena, including molecular absorption, scattering effects, and turbulence, elucidating their intricate interactions with THz waves and the resultant implications for channel modeling and system design. Furthermore, we investigate the unique security challenges posed by THz communications, examining potential vulnerabilities and proposing novel countermeasures to enhance the resilience of these high-frequency systems against eavesdropping and other security threats. Finally, we discuss the challenges and limitations of such high-frequency wireless communications and provide insights into future research prospects for realizing the 6G vision, emphasizing the need for innovative solutions to overcome the atmospheric hurdles and security concerns in THz communications.

eess.SP↗