Searcharxiv⌕ Search

arXiv subjects

Dong Wang

Publications and source records attributed to Dong Wang.

At least 73 records · Page 4Linked to original sources
arXiv2026

Deep Learning-Based Estimation of Ground Reaction Forces in Parkinsonian Gait Using an Optimized Set of IMU Data

Accurate gait analysis in Parkinson's disease (PD) typically relies on laboratory-based systems to capture biomechanical data, such as ground reaction forces (GRFs). Estimating GRFs using inertial measurement units (IMUs) provides a feasible alternative. However, this approach remains challenging in pathological gait like PD due to its high variability and complexity. Moreover, existing monitoring approaches often require multiple body-mounted sensors, which limit practicality and reduce patient compliance. To date, no study has investigated the application of deep learning approaches to address this challenge. This study proposes, for the first time, a deep learning framework to estimate bilateral vertical GRFs (vGRFs) in PD using an optimized set of wearable IMUs. A hybrid CNN-BiLSTM model was trained separately on data from 61 PD patients and 65 healthy controls (HC) using 13 IMUs. The model achieved high intra-subject accuracy ($R^2$ = 0.98) and strong inter-subject generalization ($R^2$ = 0.93 for HC, $R^2$ = 0.91 for PD). Sensor configuration was found to significantly influence estimation accuracy, with optimal sensor placement varying between PD patients and HC. For PD patients, estimation accuracy dropped markedly when reducing to a single IMU. The optimal configuration for PD used four IMUs. We identified a minimal setup with only two IMUs still enabled robust estimation. This compact setup offers a practical and scalable solution. Overall, the proposed approach supports the development of wearable vGRF-based gait analysis systems for Parkinsonian gait and potentially other pathological conditions, enabling accessible clinical assessments, remote monitoring, and personalized rehabilitation.

cs.LG↗
arXiv2026

Evaluating Fuzz Testing for Reinforcement Learning Agents

Reinforcement Learning (RL) agents are increasingly deployed in safety-critical domains such as robotics, autonomous driving, and drone control, where unexpected behaviors may lead to severe real-world consequences. Fuzz testing has recently emerged as a promising method for exploring the vast state spaces of RL agents and exposing crashes. Although numerous RL fuzzing methods have been proposed, existing studies often differ in evaluation settings, baselines, and metrics, making it difficult to draw reliable conclusions about their relative effectiveness and practical usefulness. To address this gap, we present the first comprehensive empirical study that systematically evaluates RL fuzzing methods from four complementary perspectives: effectiveness, diversity, efficiency, and practical utility. We benchmark five state-of-the-art methods alongside random testing under unified configurations across three environments of increasing complexity (MountainCar, BipedalWalker, and CARLA), and further assess the downstream usefulness of detected crashes for agent robustness improvement and safety monitoring. Our results reveal several key insights. For instance,throughput-oriented methods like MDPFuzz demonstrate superior effectiveness and efficiency in crash discovery, while methods explicitly designed to encourage exploration like SeqDivFuzz excel at uncovering diverse crash behaviors. We also show that fuzzing-generated crashes can meaningfully improve agent robustness and enable accurate safety monitoring with strong cross-method generalization. Beyond these empirical findings, we distill actionable guidance for both researchers and practitioners, highlighting the benefits of combining complementary fuzzing strategies and adopting multi-level diversity analysis to achieve more comprehensive and practical RL testing.

cs.LG↗
arXiv2026

A generalized matrix-valued Allen--Cahn model and its numerical solution

This paper introduces a generalized matrix-valued Allen--Cahn model, where the unknown matrix-valued field belongs to $\mathbb{R}^{m_1\times m_2}$ with dimension $m_1\geq m_2$. By taking different values of $m_1$ and $m_2$, this model covers the classical scalar-valued, vector-valued, and square-matrix-valued Allen--Cahn equations. At the continuous level, the proposed model is proven to admit a unique solution satisfying the maximum bound principle (MBP) and the energy dissipation law. At the discrete level, a class of arbitrarily high-order exponential time differencing Runge-Kutta (ETDRK) schemes is investigated that preserve the MBP unconditionally. Moreover, we prove that the first- and second-order ETDRK schemes satisfy the discrete energy dissipation unconditionally, while third- and higher-order schemes preserve the discrete energy dissipation under suitable time-step constraints. The proof of sharp convergence order in time is provided. Numerical experiments are carried out to confirm our theoretical results.

math.NA↗
arXiv2026

Maximal Fisher Budget Forces Blind Directions in Bipartite Collective SU(d) Metrology

For pure two-qudit probes used in a single fixed setting to estimate a common SU(d) transformation, maximal total Fisher sensitivity and full local identifiability are mutually exclusive. We derive exact identities that determine the trace of the quantum Fisher information matrix from two measurable probe properties: exchange symmetry and collective polarization. Every maximizer is exchange symmetric and maximally entangled. Its Fisher matrix has equal sensitivity in every visible direction, while $d(d-1)/2$ generator directions are blind. More generally, weak compatibility, defined by vanishing mean generator commutators, forces at least $\lfloor d/2\rfloor$ blind directions for every pure bipartite probe. A probe-defined antiunitary invariant classifies the complete Fisher spectrum at maximal entanglement, and approaches to the optimum that preserve exchange symmetry have a divergent Holevo cost. The bipartite obstruction is sharp in particle number: for $N\geq 3$, generalized GHZ probes can satisfy weak compatibility, attain the corresponding $N$-partite Fisher trace bound, and retain full local identifiability. Hence, Fisher sensitivity, weak compatibility, identifiability, and attainable precision are distinct resources in quantum metrology with noncommuting generators.

quant-ph↗
arXiv2026

Context Matters: Improving the Practical Reliability of LLM-Based Unit Test Generation

Automated unit test generation has recently benefited from advances in large language models (LLMs), yet our industrial deployments reveal a persistent gap between promising research results and practical usability. In real-world projects with complex frameworks and cross-file dependencies, LLM-generated tests frequently fail to compile, require costly manual repair, or provide unstable coverage improvements. This paper reports our experience in designing, deploying, and evaluating CATGen, a context-aware workflow for LLM-based unit test generation, informed by repeated industrial failures and refinements. Rather than relying on LLMs to infer incomplete project context, we found that compilation robustness critically depends on making project-level dependencies explicit, stabilizing test class scaffolding, and replacing iterative LLM-based repair with lightweight static analysis. These experience-driven insights shaped CATGen's multi-stage design, which combines structured context retrieval, deterministic test skeleton construction, and program analysis-based post-processing. We evaluate CATGen on real-world complex focal methods from proprietary industrial projects and additionally on the Defects4J benchmark to assess generalizability. Across both settings, CATGen substantially improves compilation success and structural coverage while significantly reducing generation time and token consumption compared to existing LLM-based approaches. Our results demonstrate that reliable LLM-based unit test generation in practice depends less on prompt engineering alone and more on systematic engineering support grounded in real-world development constraints.

cs.SE↗
arXiv2026

Dr. Zero: Self-Evolving Search Agents without Training Data

As high-quality data becomes increasingly difficult to obtain, self-evolution without curated training data has emerged as a promising paradigm. This approach allows large language models (LLMs) to autonomously generate and solve complex problems, thereby improving their reasoning capabilities. However, multi-turn search agents struggle in this setting due to limited question diversity and the substantial compute required for multi-step reasoning and tool use. In this work, we introduce Dr. Zero, a framework that enables search agents to effectively self-evolve without human-annotated training data, relying solely on an external search engine as their knowledge environment. In particular, we design a self-evolution feedback loop where a proposer generates structurally diverse questions to train a solver initialized from the same base model. As the solver evolves, it incentivizes the proposer to produce increasingly difficult yet solvable tasks, thus establishing an automated curriculum to refine both agents. To enhance training efficiency, we also introduce hop-grouped relative policy optimization (HRPO). This method clusters structurally similar questions to construct group-level baselines, effectively minimizing the sampling overhead in evaluating each query's individual difficulty and solvability. Consequently, HRPO significantly reduces the compute requirements for proposer training and reward estimation without compromising performance or stability. Extensive experimental results demonstrate that Dr. Zero matches or surpasses fully supervised search agents on several question answering benchmarks, showing that strong agentic search and evidence-grounded reasoning can emerge solely through self-evolution.

cs.AI↗
arXiv2026

Topology-Aware Tokenization for Generative Recommendation

Generative recommendation reformulates sequential recommendation as an autoregressive generation task, yet a critical issue in this paradigm remains overlooked: topology distortion in item tokenization. In particular, we observe that the intrinsic adjacency relationships of items in the pretrained semantic embedding space are significantly disrupted after quantization. This topology distortion misleads the model's perception of item similarity, ultimately bottlenecking the accuracy of generative recommendations. To address this issue, we propose Topology-Aware Tokenization (TopoTok), an item tokenization framework that preserves item relational structure throughout the quantization hierarchy. Different from the prior monolithic supervision in tokenization, TopoTok introduces a multi-level distillation scheme to progressively recover the topology from coarse to fine granularity: 1) Inter-Group Distillation to capture global cluster-wise relations; 2) Intra-Group Distillation to refine local structures within semantic clusters; and 3) Inter-Item Distillation to enforce fine-grained alignment at the individual item level. Extensive experiments on three benchmark datasets demonstrate that TopoTok effectively alleviates topology distortion and consistently outperforms state-of-the-art tokenizers, achieving significant performance gains of up to 9.42% in Recall@5.

cs.IR↗
arXiv2026

KernelDiag: Agent-Based Root Cause Diagnosis for Kernel Crashes

The Linux kernel is one of the most complex software systems, where automated fuzzing continuously exposes thousands of crashes, yet root-cause diagnosis remains a manual and time-consuming bottleneck. Existing LLM-based root cause analysis (RCA) techniques, effective for distributed systems, do not readily generalize to kernel debugging due to sparse low-level artifacts, heterogeneous diagnostic evidence (e.g., syscalls, logs, and crash reports), and complex non-linear fault propagation that demands fine-grained method-level reasoning. To address these challenges, we propose KernelDiag, an agent-based framework for kernel root-cause diagnosis via structured causal reasoning. KernelDiag first aligns heterogeneous diagnostic artifacts through log-to-code mapping, and then employs artifact-specialized agents to iteratively reason over source-level program semantics and crash-specific kernel configurations. The inferred causal dependencies are incrementally organized into structured Evidence Graphs, enabling accurate faulty-method localization and causal explanations. We evaluate KernelDiag on the real-world KGYM benchmark. KernelDiag consistently outperforms state-of-the-art localization approaches at both file and method levels, achieving significant improvements in Top@k accuracy, including up to 4x and 2x gains in challenging settings without explicit hints. Furthermore, both human and LLM-assisted evaluations show that KernelDiag generates accurate, coherent, and actionable diagnostic explanations. Overall, this work lays the foundation for automated kernel root-cause diagnosis by bridging low-level diagnostic evidence with source-level causal reasoning.

cs.SE↗
arXiv2026

Global well-posedness of strong solutions to a model for the morning glory cloud

In this paper, we investigate the global well-posedness of strong solution to a model recently derived by Constantin-Johnson \cite{CaJo} which describes the nonlinear wave propagation in the troposphere, especially for the morning glory cloud. Assuming that the initial velocity $v_0\in H^1$ and the thermodynamic forcing term $K\in L^2(0,T;L^2)$, we show that there exists a unique global strong solution to the initial boundary value problem of this system. Similar result was only known before for sufficiently small initial data in the existing literature.

math.AP↗
arXiv2026

Q-GeoMem: Question-Guided Geometric Memory for Video Spatial Reasoning

Video spatial reasoning requires accumulating viewpoint-dependent evidence over time while retaining information useful to the question being asked. Existing spatial video-language models improve geometric perception and long-range context modeling, but often treat memory as a generic temporal cache, which can introduce redundant or irrelevant evidence and weaken long-horizon reasoning. We propose Q-GeoMem, a question-guided geometric memory framework for video spatial reasoning. Q-GeoMem injects camera-conditioned geometry into visual tokens and maintains two complementary memories: a Fine-Grained Context Bank for recent dense features and camera states, and a Semantic-Geometric Evidence Bank for compact long-range evidence. For each candidate frame, a calibrated Q-Former estimates question relevance, while novelty and evidence utility are recomputed with respect to the active evidence bank. The resulting relevance-novelty utility controls capacity-based replacement and serves as an attention bias during memory reading. During reasoning, both memories are read before update and adaptively fused with the current frame representation. Extensive experiments across two in-domain and five out-of-distribution benchmarks, and controlled memory analyses show that Q-GeoMem achieves state-of-the-art performance in the evaluated settings and validate the effectiveness of question-guided geometric evidence selection.

cs.CV↗
arXiv2026

A CubeSat Electronics System for Dual-Satellite Coordinated Soft X-Ray Polarimetry

Addressing the unique requirements for a wide field of view and rapid response in soft X-ray polarization measurements of transient sources such as gamma-ray bursts, this paper proposes the design of a high-reliability CubeSat payload electronics system for dual-satellite cooperative observation. The system employs a Gas Microchannel Pixel Detector (GMPD) and a Topmetal-L sensor as its core components, forming a highly integrated, low-noise payload hardware platform. It achieves a field of view of 90$^\circ$ $\times$ 90$^\circ$, a sensitive area of 3.69 cm$^{2}$, and a power consumption of less than 6 W, operating within an energy range of 2--10 keV. The system incorporates autonomous high-voltage (HV) ramp-up/ramp-down control and a dual-protection mechanism based on count rate and discharge events, enabling in-orbit responses to risks such as the South Atlantic Anomaly and solar particle events. It also supports single-event upset detection and recovery, as well as in-orbit firmware upgrades. The communication interface adopts a redundant primary/backup Controller Area Network (CAN) bus design, with measured channel switching times of less than 100 ms, meeting the demands for real-time command interaction in dual-satellite coordination. By utilizing a large-array pixel sensor with region-of-interest readout and integrating an in-orbit track compression algorithm, the system significantly reduces data storage and downlink transmission resource burdens. Ground tests demonstrate an equivalent noise charge of 22.35 e$^{-}$, HV monitoring linearity better than $\pm$0.5\%, and an output range extending to -5 kV. Thermal vacuum cycling tests show no performance degradation after five cycles between -5 $^\circ$C and 40 $^\circ$C. This work demonstrates the system's capability for autonomous observation, intelligent coordination, and reliable operation in complex space environments.

physics.ins-det↗
arXiv2026

Large Language Models for Multi-Lingual Equivalent Mutant Detection: An Extended Empirical Study

Mutation testing is a powerful technique for ensuring software quality. However, the presence of equivalent mutants introduces unnecessary costs and biases, limiting its practical effectiveness. Although numerous equivalent mutant detection (EMD) methods have been proposed, they often face distinct challenges: pure-code analysis methods can be limited by their reliance on specific compiler infrastructures, while existing machine-learning approaches remain constrained by scarce training data and limited generalization to unseen mutants. Large language models (LLMs) have recently demonstrated remarkable performance across diverse code-related tasks by better capturing program semantics. Yet their potential for EMD remains largely unexplored, particularly in the multi-lingual context. This paper presents the first comprehensive empirical study on LLMs for EMD, using 3,302 Java and 1,088 C mutant pairs to benchmark against state-of-the-art methods, explore strategy variations, assess efficiency, and evaluate cross-lingual generalization. Experimental results show that LLM-based approaches achieve higher F1-scores than the evaluated traditional methods, with fine-tuned code embedding yielding the highest detection accuracy among the tested strategies. Moreover, LLM-based approaches strike a practical balance between effectiveness and efficiency with inference times comparable to existing machine-learning models. Importantly, fine-tuned LLMs demonstrate measurable generalization across programming languages. These findings establish LLMs as a viable and efficient approach for tackling the longstanding challenge of equivalent mutant detection, offering new directions for advancing mutation testing in practice.

cs.SE↗
arXiv2026

Learn Weightlessness: Imitate Non-Self-Stabilizing Motions on Humanoid Robot

The integration of imitation and reinforcement learning has enabled remarkable advances in humanoid whole-body control, facilitating diverse human-like behaviors. However, research on environment-dependent motions remains limited. Existing methods typically enforce rigid trajectory tracking while neglecting physical interactions with the environment. We observe that humans naturally exploit a "weightless" state during non-self-stabilizing (NSS) motions--selectively relaxing specific joints to allow passive body--environment contact, thereby stabilizing the body and completing the motion. Inspired by this biological mechanism, we design a weightlessness-state auto-labeling strategy for dataset annotation; and we propose the Weightlessness Mechanism (WM), a method that dynamically determines which joints to relax and to what level, together enabling effective environmental interaction while executing target motions. We evaluate our approach on 3 representative NSS tasks: sitting on chairs of varying heights, lying down on beds with different inclinations, and leaning against walls via shoulder or elbow. Extensive experiments in simulation and on the Unitree G1 robot demonstrate that our WM method, trained on single-action demonstrations without any task-specific tuning, achieves strong generalization across diverse environmental configurations while maintaining motion stability. Our work bridges the gap between precise trajectory tracking and adaptive environmental interaction, offering a biologically-inspired solution for contact-rich humanoid control.

cs.RO↗
arXiv2026

RPG: Robust Policy Gating for Smooth Multi-Skill Transitions in Humanoid Fighting

Humanoid robots have demonstrated impressive motor skills in a wide range of tasks, yet whole-body control for humanlike long-time, dynamic fighting remains particularly challenging due to the stringent requirements on agility and stability. While imitation learning enables robots to execute human-like fighting skills, existing approaches often rely on switching among multiple single-skill policies or employing a general policy to imitate input reference motions. These strategies suffer from instability when transitioning between skills, as the mismatch of initial and terminal states across skills or reference motions introduces out-of-domain disturbances, resulting in unsmooth or unstable behaviors. In this work, we propose RPG, a hybrid expert policy framework, for smooth and stable humanoid multi-skills transition. Our approach incorporates motion transition randomization and temporal randomization to train a unified policy that generates agile fighting actions with stability and smoothness during skill transitions. Furthermore, we design a control pipeline that integrates walking/running locomotion with fighting skills, allowing humanlike long-time combat of arbitrary duration that can be seamlessly interrupted or transit action policies at any time. Extensive experiments in simulation demonstrate the effectiveness of the proposed framework, and real-world deployment on the Unitree G1 humanoid robot further validates its robustness and applicability.

cs.RO↗
arXiv2026

DLR: Zero-Inference-Cost Latent Residuals for Low-Rank Pre-Training

Large language models have driven recent progress in language and multimodal AI, yet pre-training them at scale is prohibitively expensive. Low-rank pre-training, which factorizes each weight matrix into a rank-r product to reduce both parameters and FLOPs, is a promising response but typically lags full-rank training in quality. We propose Duplicated Latent Residual (DLR), a training-only, parameter-free, foldable plug-in for low-rank pre-training. DLR augments the standard low-rank output Bz with a fixed structured residual alpha/sqrt(K) * Expand_K(z) that replicates each latent coordinate K = ceil(d_out/r) times across the output. With alpha fixed, DLR adds zero learnable parameters per layer; after training, it is absorbed into the up-projection in closed form, B* = B + alpha/sqrt(K) R^T, so deployment parameter count, FLOPs and memory match the underlying low-rank backbone exactly. Across LLaMA models from 60M to 7B parameters, DLR strengthens low-rank pre-training on C4 validation perplexity in most settings, with the clearest gains at 130M and above; folded checkpoints transfer cleanly to supervised fine-tuning on standard benchmarks.

cs.LG↗
arXiv2026

A fast sum-of-Gaussians algorithm for the high-dimensional fractional Fokker-Planck equation

We present a fast, high-order algorithm for the free-space fractional Fokker-Planck equation (FFPE) in arbitrary spatial dimension. Its fundamental solution, corresponding to a Dirac-delta initial condition, is obtained from the explicit Fourier representation by applying a sum-of-Gaussians (SOG) approximation to the nonseparable stretched exponential, using its complete monotonicity as the Laplace transform of a one-sided $α$-stable density. Each Gaussian term is an ordinary heat kernel and therefore factorizes across spatial coordinates. On a tensor-product grid, the separated form can be assembled in $O(MdN)$ work and storage, rather than forming all $O(N^d)$ grid values, where $M$ is the number of Gaussian terms and $N$ is the number of points per dimension. We prove an a~priori error estimate for the pure-fractional fundamental solution and give a parameter-selection procedure for prescribed accuracy over specified ranges of space and time. In numerical experiments the method achieves more than ten digits of relative accuracy, with $M$ growing only logarithmically in the inverse tolerance, and maintains this accuracy in dimensions up to $d=10^{5}$. This exceeds the dimensions reached in comparable radial-quadrature tests, where the integrand becomes increasingly oscillatory as the dimension grows. Because the method represents the fundamental solution as a separated sum of heat kernels, any initial datum given as a finite sum of tensor products can be evolved in closed form using only one-dimensional convolutions. This yields a computable class of high-dimensional solutions that is amenable to error analysis, and tensor neural networks provide one possible way to construct such separated representations for more general data.

math.NA↗
arXiv2026

ZeroWBC: Learning Natural Whole-Body Humanoid Interaction from Human Egocentric Data

Achieving versatile and natural whole-body humanoid interaction control remains challenging due to the high cost of whole-body teleoperation data. We present ZeroWBC, a teleoperation-free framework that learns humanoid whole-body interaction from human egocentric videos paired with synchronized whole-body motion and text annotations. ZeroWBC adopts a generation-then-tracking formulation to tackle the static scene whole-body interaction control problem. Given an initial egocentric image and a language instruction, a fine-tuned Vision-Language Model generates future human whole-body motion tokens, which are decoded into continuous motions and retargeted to the humanoid. The resulting reference motions, together with root and key body-part trajectories, are then executed by a general interactive motion tracking policy. To improve interaction performance, we introduce an interaction-oriented tracking reward that prioritizes global root and key body-part trajectory alignment while preserving natural whole-body motion. Experiments on the Unitree G1 humanoid robot show that ZeroWBC enables diverse scene-aware behaviors without robot teleoperation demonstrations. These results suggest a scalable paradigm for learning natural humanoid whole-body interaction from human egocentric data.

cs.RO↗
arXiv2026

Quantifying and Probing Multipartite Entanglement via Minimum Entanglement Drop

Quantifying genuine multipartite entanglement remains a significant challenge. We propose a multipartite entanglement monotone defined by the minimum entanglement drop -- the reduction in global one-to-group entanglement upon tracing out a single particle. We formulate a computationally efficient variant using tangle and negativity to ensure non-vanishing values for W-class states, and rigorously prove it is a valid monotone under local operations and classical communication. In the tripartite regime, the minimum tangle drop is physically equivalent to the minimum pairwise concurrence. We establish an operational framework where the entanglement drop acts as a structural probe: by assessing sensitivity to qubit loss, it identifies inseparable clusters, extracting connectivity fingerprints that uniquely differentiate graph topologies within the same local Clifford equivalence class. Integrating this mapping with classical shadows enables efficient experimental estimation and dynamic tracking of entanglement network evolution. We derive exact analytical solutions for n-qubit W states under environmental noise, revealing robust scaling behaviors. Finally, we acknowledge limitations, noting that diagnostic sensitivity strictly vanishes for highly robust states such as the 5-qubit error-correcting code.

quant-ph↗