arXiv subjects

Zheng Li

Publications and source records attributed to Zheng Li.

At least 145 records · Page 8Linked to original sources

arXiv2025

VLCache: Computing 2% Vision Tokens and Reusing 98% for Vision-Language Inference

This paper presents VLCache, a cache reuse framework that exploits both Key-Value (KV) cache and encoder cache from prior multimodal inputs to eliminate costly recomputation when the same multimodal inputs recur. Unlike previous heuristic approaches, we formally identify the cumulative reuse error effect and demonstrate how to minimize the non-prefix cache reuse error effectively. We further analyze the varying importance of model layers and propose a dynamic, layer-aware recomputation strategy to balance accuracy and efficiency. Experimental results show that VLCache achieves an accuracy on par with full recomputation, while requiring only 2-5% of the tokens to compute, yielding 1.2x-16x TTFT speedups. We develop an experimental implementation of the proposed VLCache pipeline based on SGLang, enabling significantly faster inference in practical deployments.

cs.CV↗

arXiv2025

Anomalous Hall effect and rich magnetic phase diagram of Mn$_{100-x}$Rh$_{x}$ epitaxial films

A series of Mn$_{100-x}$Rh$_x$ ($20 \le x \le 50$) thin films were epitaxially grown on the MgO substrate using magnetron sputtering technique, and were systematically investigated by magnetization, longitudinal electrical resistivity, and transverse Hall resistivity. After optimizing the growth conditions, phase-pure Mn$_{100-x}$Rh$_x$ films with a cubic CsCl-type structure were obtained, and their magnetic phase diagram was built. The manipulation of Rh content leads to a rich magnetic phase diagram, where three different regimes can be identified: for $x < 40$, Mn$_{100-x}$Rh$_x$ films undergo a ferromagnetic (FM) transition below $T_\mathrm{C} \approx$ 330-350 K; for $40 \le x \le 45$, in addition to the FM transition at $T_\mathrm{C} \approx$ 200 K, Mn$_{100-x}$Rh$_x$ films undergo a FM-to-antiferromagnetic (AFM) transition at $T_\mathrm{N} \approx$ 120 K; finally for $x > 45$, only one AFM transition at $T_\mathrm{N} \approx$ 150 K can be tracked. All the Mn$_{100-x}$Rh$_x$ films exhibit distinct anomalous Hall effect in their magnetically ordered state, which is most likely due to the intrinsic Berry-curvature mechanism. In addition, all the anomalous Hall transport properties, including the resistivity, conductivity, and angle exhibit a strong correlation with the magnetic properties of Mn$_{100-x}$Rh$_x$ films, which become most evident for $x$ = 35. Our systematic investigations suggest a strong correlation between magnetic properties and electronic band topology in Mn$_{100-x}$Rh$_x$ films, highlighting their great potential for AFM spintronics.

cond-mat.str-el↗

arXiv2025

Estimation of Treatment Effects based on Kernel Matching

Kernel matching is a widely used technique for estimating treatment effects, particularly valuable in observational studies where randomized controlled trials are not feasible. While kernel-matching approaches have demonstrated practical advantages in exploiting similarities between treated and control units, their theoretical properties have remained only partially explored. In this paper, we make a key contribution by establishing the asymptotic normality and consistency of kernel-matching estimators for both the average treatment effect (ATE) and the average treatment effect on the treated (ATT) through influence function techniques, thereby providing a rigorous theoretical foundation for their use in causal inference. Furthermore, we derive the asymptotic distributions of the ATE and ATT estimators when the propensity score is estimated rather than known, extending the theoretical guarantees to the practically relevant cases. Through extensive Monte Carlo simulations, the estimators exhibit consistently improved performance over standard treatment-effect estimators. We further illustrate the method by analyzing the National Supported Work Demonstration job-training data with the kernel-matching estimator.

stat.ME↗

arXiv2025

Can VLMs Detect and Localize Fine-Grained AI-Edited Images?

Fine-grained detection and localization of localized image edits is crucial for assessing content authenticity, especially as modern diffusion models and image editors can produce highly realistic manipulations. However, this problem faces three key challenges: (1) most AIGC detectors produce only a global real-or-fake label without indicating where edits occur; (2) traditional computer vision methods for edit localization typically rely on costly pixel-level annotations; and (3) there is no large-scale, modern benchmark specifically targeting edited-image detection. To address these gaps, we develop an automated data-generation pipeline and construct FragFake, a large-scale benchmark of AI-edited images spanning multiple source datasets, diverse editing models, and several common edit types. Building on FragFake, we are the first to systematically study vision language models (VLMs) for edited-image classification and edited-region localization. Our experiments show that pretrained VLMs, including GPT4o, perform poorly on this task, whereas fine-tuned models such as Qwen2.5-VL achieve high accuracy and substantially higher object precision across all settings. We further explore GRPO-based RLVR training, which yields modest metric gains while improving the interpretability of model outputs. Ablation and transfer analyses reveal how data balancing, training size, LoRA rank, and training domain affect performance, and highlight both the potential and the limitations of cross-editor and cross-dataset generalization. We anticipate that this work will establish a solid foundation to facilitate and inspire subsequent research endeavors in the domain of multimodal content authenticity.

cs.CV↗

arXiv2025

MBFL-DKMR: Improving Mutation-based Fault Localization through Denoising-based Kill Matrix Refinement

Software debugging is a critical and time-consuming aspect of software development, with fault localization being a fundamental step that significantly impacts debugging efficiency. Mutation-Based Fault Localization (MBFL) has gained prominence due to its robust theoretical foundations and fine-grained analysis capabilities. However, recent studies have identified a critical challenge: noise phenomena, specifically the false kill relationships between mutants and tests, which significantly degrade localization effectiveness. While several approaches have been proposed to rectify the final localization results, they do not directly address the underlying noise. In this paper, we propose a novel approach to refine the kill matrix, a core data structure capturing mutant-test relationships in MBFL, by treating it as a signal that contains both meaningful fault-related patterns and high-frequency noise. Inspired by signal processing theory, we introduce DKMR (Denoising-based Kill Matrix Refinement), which employs two key stages: (1) signal enhancement through hybrid matrix construction to improve the signal-to-noise ratio for better denoising, and (2) signal denoising via frequency domain filtering to suppress noise while preserving fault-related patterns. Building on this foundation, we develop MBFL-DKMR, a fault localization framework that utilizes the refined matrix with fuzzy values for suspiciousness calculation. Our evaluation on Defects4J v2.0.0 demonstrates that MBFL-DKMR effectively mitigates the noise and outperforms the state-of-the-art MBFL techniques. Specifically, MBFL-DKMR achieves 129 faults localized at Top-1 compared to 85 for BLMu and 103 for Delta4Ms, with negligible additional computational overhead (0.11 seconds, 0.001\% of total time). This work highlights the potential of signal processing techniques to enhance the effectiveness of MBFL by refining the kill matrix.

cs.SE↗

arXiv2025

FLIMs: Fault Localization Interference Mutants, Definition, Recognition and Mitigation

Mutation-based Fault Localization (MBFL) has been widely explored for automated software debugging, leveraging artificial mutants to identify faulty code entities. However, MBFL faces significant challenges due to interference mutants generated from non-faulty code entities but can be killed by failing tests. These mutants mimic the test sensitivity behaviors of real faulty code entities and weaken the effectiveness of fault localization. To address this challenge, we introduce the concept of Fault Localization Interference Mutants (FLIMs) and conduct a theoretical analysis based on the Reachability, Infection, Propagation, and Revealability (RIPR) model, identifying four distinct interference causes. Building on this, we propose a novel approach to semantically recognize and mitigate FLIMs using LLM-based semantic analysis, enhanced by fine-tuning techniques and confidence estimation strategies to address LLM output instability. The recognized FLIMs are then mitigated by refining the suspiciousness scores calculated from MBFL techniques. We integrate FLIM recognition and mitigation into the MBFL workflow, developing MBFL-FLIM, a fault localization framework that enhances MBFL's effectiveness by reducing misleading interference while preserving real fault-revealing information. Our empirical experiments on the Defects4J benchmark with 395 program versions using eight LLMs demonstrate MBFL-FLIM's superiority over traditional SBFL and MBFL methods, advanced dynamic feature-based approaches, and recent LLM-based fault localization techniques. Specifically, MBFL-FLIM achieves an average improvement of 44 faults in the Top-1 metric, representing a significant enhancement over baseline methods. Further evaluation confirms MBFL-FLIM's robust performance in multi-fault scenarios, with ablation experiments validating the contributions of the fine-tuning and confidence estimation components.

cs.SE↗

arXiv2025

AnchorOPT: Towards Optimizing Dynamic Anchors for Adaptive Prompt Learning

Existing prompt learning methods, which are built upon CLIP models, leverage textual tokens as anchors to guide the learnable soft tokens. This guidance improves CLIP generalizations. However, these anchors-static in both value and position-lack cross-task and stage-adaptive flexibility. To address this limitation, we propose AnchorOPT, a dynamic anchor-based prompt learning framework. Specifically, AnchorOPT introduces dynamism in two key dimensions: (i) anchor values eschew handcrafted explicit textual tokens (e.g., "shape", "color"), instead learning dynamically from task-specific data; and (ii) the positional relationship between anchor and soft tokens is no longer fixed but adaptively optimized via a learnable position matrix conditioned on the training stage and task context. Training occurs in two stages: we first learn the anchor tokens, then freeze and transfer them to the second stage for optimization of soft tokens and the position matrix. Extensive experiments demonstrate that using only a simple learnable anchor and position matrix achieves performance comparable to or exceeding some methods incorporating additional learnable modules or regularization techniques. As a plug-and-play module, AnchorOPT integrates seamlessly into existing frameworks, yielding consistent performance gains across diverse datasets. Code is publicly available at https://github.com/zhengli97/ATPrompt.

cs.CV↗

arXiv2025

Adaptive Knowledge Transfer for Cross-Disciplinary Cold-Start Knowledge Tracing

Cross-Disciplinary Cold-start Knowledge Tracing (CDCKT) faces a critical challenge: insufficient student interaction data in the target discipline prevents effective knowledge state modeling and performance prediction. Existing cross-disciplinary methods rely on overlapping entities between disciplines for knowledge transfer through simple mapping functions, but suffer from two key limitations: (1) overlapping entities are scarce in real-world scenarios, and (2) simple mappings inadequately capture cross-disciplinary knowledge complexity. To overcome these challenges, we propose Mixed of Experts and Adversarial Generative Network-based Cross-disciplinary Cold-start Knowledge Tracing Framework. Our approach consists of three key components: First, we pre-train a source discipline model and cluster student knowledge states into K categories. Second, these cluster attributes guide a mixture-of-experts network through a gating mechanism, serving as a cross-domain mapping bridge. Third, an adversarial discriminator enforces feature separation by pulling same-attribute student features closer while pushing different-attribute features apart, effectively mitigating small-sample limitations. We validate our method's effectiveness across 20 extreme cross-disciplinary cold-start scenarios.

cs.IR↗

arXiv2025

UltraDP: Generalizable Carotid Ultrasound Scanning with Force-Aware Diffusion Policy

Ultrasound scanning is a critical imaging technique for real-time, non-invasive diagnostics. However, variations in patient anatomy and complex human-in-the-loop interactions pose significant challenges for autonomous robotic scanning. Existing ultrasound scanning robots are commonly limited to relatively low generalization and inefficient data utilization. To overcome these limitations, we present UltraDP, a Diffusion-Policy-based method that receives multi-sensory inputs (ultrasound images, wrist camera images, contact wrench, and probe pose) and generates actions that are fit for multi-modal action distributions in autonomous ultrasound scanning of carotid artery. We propose a specialized guidance module to enable the policy to output actions that center the artery in ultrasound images. To ensure stable contact and safe interaction between the robot and the human subject, a hybrid force-impedance controller is utilized to drive the robot to track such trajectories. Also, we have built a large-scale training dataset for carotid scanning comprising 210 scans with 460k sample pairs from 21 volunteers of both genders. By exploring our guidance module and DP's strong generalization ability, UltraDP achieves a 95% success rate in transverse scanning on previously unseen subjects, demonstrating its effectiveness.

cs.RO↗

arXiv2025

BudgetLeak: Membership Inference Attacks on RAG Systems via the Generation Budget Side Channel

Retrieval-Augmented Generation (RAG) enhances large language models by integrating external knowledge, but reliance on proprietary or sensitive corpora poses various data risks, including privacy leakage and unauthorized data usage. Membership inference attacks (MIAs) are a common technique to assess such risks, yet existing approaches underperform in RAG due to black-box constraints and the absence of strong membership signals. In this paper, we identify a previously unexplored side channel in RAG systems: the generation budget, which controls the maximum number of tokens allowed in a generated response. Varying this budget reveals observable behavioral patterns between member and non-member queries, as members gain quality more rapidly with larger budgets. Building on this insight, we propose BudgetLeak, a novel membership inference attack that probes responses under different budgets and analyzes metric evolution via sequence modeling or clustering. Extensive experiments across four datasets, three LLM generators, and two retrievers demonstrate that BudgetLeak consistently outperforms existing baselines, while maintaining high efficiency and practical viability. Our findings reveal a previously overlooked data risk in RAG systems and highlight the need for new defenses.

cs.CR↗

arXiv2025

CATS-V2V: A Real-World Vehicle-to-Vehicle Cooperative Perception Dataset with Complex Adverse Traffic Scenarios

Vehicle-to-Vehicle (V2V) cooperative perception has great potential to enhance autonomous driving performance by overcoming perception limitations in complex adverse traffic scenarios (CATS). Meanwhile, data serves as the fundamental infrastructure for modern autonomous driving AI. However, due to stringent data collection requirements, existing datasets focus primarily on ordinary traffic scenarios, constraining the benefits of cooperative perception. To address this challenge, we introduce CATS-V2V, the first-of-its-kind real-world dataset for V2V cooperative perception under complex adverse traffic scenarios. The dataset was collected by two hardware time-synchronized vehicles, covering 10 weather and lighting conditions across 10 diverse locations. The 100-clip dataset includes 60K frames of 10 Hz LiDAR point clouds and 1.26M multi-view 30 Hz camera images, along with 750K anonymized yet high-precision RTK-fixed GNSS and IMU records. Correspondingly, we provide time-consistent 3D bounding box annotations for objects, as well as static scenes to construct a 4D BEV representation. On this basis, we propose a target-based temporal alignment method, ensuring that all objects are precisely aligned across all sensor modalities. We hope that CATS-V2V, the largest-scale, most supportive, and highest-quality dataset of its kind to date, will benefit the autonomous driving community in related tasks.

cs.CV↗

arXiv2025

Compensating Distribution Drifts in Class-incremental Learning of Pre-trained Vision Transformers

Recent advances have shown that sequential fine-tuning (SeqFT) of pre-trained vision transformers (ViTs), followed by classifier refinement using approximate distributions of class features, can be an effective strategy for class-incremental learning (CIL). However, this approach is susceptible to distribution drift, caused by the sequential optimization of shared backbone parameters. This results in a mismatch between the distributions of the previously learned classes and that of the updater model, ultimately degrading the effectiveness of classifier performance over time. To address this issue, we introduce a latent space transition operator and propose Sequential Learning with Drift Compensation (SLDC). SLDC aims to align feature distributions across tasks to mitigate the impact of drift. First, we present a linear variant of SLDC, which learns a linear operator by solving a regularized least-squares problem that maps features before and after fine-tuning. Next, we extend this with a weakly nonlinear SLDC variant, which assumes that the ideal transition operator lies between purely linear and fully nonlinear transformations. This is implemented using learnable, weakly nonlinear mappings that balance flexibility and generalization. To further reduce representation drift, we apply knowledge distillation (KD) in both algorithmic variants. Extensive experiments on standard CIL benchmarks demonstrate that SLDC significantly improves the performance of SeqFT. Notably, by combining KD to address representation drift with SLDC to compensate distribution drift, SeqFT achieves performance comparable to joint training across all evaluated datasets. Code: https://github.com/raoxuan98-hash/sldc.git.

cs.CV↗

arXiv2025

A High-Throughput GPU Framework for Adaptive Lossless Compression of Floating-Point Data

The torrential influx of floating-point data from domains like IoT and HPC necessitates high-performance lossless compression to mitigate storage costs while preserving absolute data fidelity. Leveraging GPU parallelism for this task presents significant challenges, including bottlenecks in heterogeneous data movement, complexities in executing precision-preserving conversions, and performance degradation due to anomaly-induced sparsity. To address these challenges, this paper introduces a novel GPU-based framework for floating-point adaptive lossless compression. The proposed solution employs three key innovations: a lightweight asynchronous pipeline that effectively hides I/O latency during CPU-GPU data transfer; a fast and theoretically guaranteed float-to-integer conversion method that eliminates errors inherent in floating-point arithmetic; and an adaptive sparse bit-plane encoding strategy that mitigates the sparsity caused by outliers. Extensive experiments on 12 diverse datasets demonstrate that the proposed framework significantly outperforms state-of-the-art competitors, achieving an average compression ratio of 0.299 (a 9.1% relative improvement over the best competitor), an average compression throughput of 10.82 GB/s (2.4x higher), and an average decompression throughput of 12.32 GB/s (2.4x higher).

cs.DB↗

arXiv2025

Absence of magnetic order and magnetic fluctuations in RuO$_{2}$

A novel magnetic class blending ferromagnetism and antiferromagnetism, termed altermagnetism, has gained significant attention for its staggered order in coordinate and momentum spaces, time-reversal symmetry-breaking phenomena, and promising applications in spintronics. Ruthenium dioxide (RuO$_{2}$) has been considered a candidate material for altermagnetism, yet the presence of magnetic moments on Ru atoms remains a subject of debate. In this study, we systematically investigated the magnetic properties of RuO$_{2}$ powder using nuclear quadrupole resonance (NQR) measurements. The NQR spectra show that there is no internal magnetic field. Furthermore, the temperature independence of spin-lattice relaxation rate, $1/T_1T$, proves that there are no magnetic fluctuations. Our results unambiguously demonstrate that Ru atoms in RuO$_{2}$ possess neither static magnetic moments nor fluctuating magnetic moments, and thus RuO$_{2}$ does not possess the magnetic characteristics essential for altermagnetism.

cond-mat.str-el↗

arXiv2025

Benchmarking Tesla's Traffic Light and Stop Sign Control: Field Dataset and Behavior Insights

Understanding how Advanced Driver-Assistance Systems (ADAS) interact with Traffic Control Devices (TCDs) is critical for assessing their influence on traffic operations, yet this interaction has received little focused empirical study. This paper presents a field dataset and behavioral analysis of Tesla's Traffic Light and Stop Sign Control (TLSSC), a mature ADAS that perceives traffic lights and stop signs. We design and execute experiments across varied speed limits and TCD types, collecting synchronized high-resolution vehicle trajectory data and driver-perspective video. From these data, we develop a taxonomy of TLSSC-TCD interaction behaviors (i.e., stopping, accelerating, and car following) and calibrate the Full Velocity Difference Model (FVDM) to quantitatively characterize each behavior mode. A novel empirical insight is the identification of a car-following threshold (~90 m). Calibration results reveal that stopping behavior is driven by strong responsiveness to both desired speed deviation and relative speed, whereas accelerating behavior is more conservative. Intersection car-following behavior exhibits smoother dynamics and tighter headways compared to standard car-following behaviors. The established dataset, behavior definitions, and model characterizations together provide a foundation for future simulation, safety evaluation, and design of ADAS-TCD interaction logic. Our dataset is available at GitHub.

cs.RO↗

arXiv2025

Local Learning for Covariate Selection in Nonparametric Causal Effect Estimation with Latent Variables

Estimating causal effects from nonexperimental data is a fundamental problem in many fields of science. A key component of this task is selecting an appropriate set of covariates for confounding adjustment to avoid bias. Most existing methods for covariate selection often assume the absence of latent variables and rely on learning the global network structure among variables. However, identifying the global structure can be unnecessary and inefficient, especially when our primary interest lies in estimating the effect of a treatment variable on an outcome variable. To address this limitation, we propose a novel local learning approach for covariate selection in nonparametric causal effect estimation, which accounts for the presence of latent variables. Our approach leverages testable independence and dependence relationships among observed variables to identify a valid adjustment set for a target causal relationship, ensuring both soundness and completeness under standard assumptions. We validate the effectiveness of our algorithm through extensive experiments on both synthetic and real-world data.

cs.LG↗

arXiv2025

Self-Rewarding PPO: Aligning Large Language Models with Demonstrations Only

Supervised fine-tuning (SFT) has emerged as a crucial method for aligning large language models (LLMs) with human-annotated demonstrations. However, SFT, being an off-policy approach similar to behavior cloning, often struggles with overfitting and poor out-of-domain generalization, especially in limited-data scenarios. To address these limitations, we propose Self-Rewarding PPO, a novel fine-tuning method that leverages on-policy techniques to enhance generalization performance. Our approach combines the strengths of SFT and proximal policy optimization (PPO) to achieve more effective alignment from demonstration data. At its core is a reward function designed as the log policy ratio between the SFT model and the pretrained base model. This function serves as an implicit reward signal, using the pretrained policy as a baseline and the SFT policy as a target. By doing so, it enables on-policy fine-tuning without relying on human preference annotations. The integration of this self-rewarding mechanism with PPO addresses key limitations of SFT, improving generalization, data efficiency, and robustness. Our empirical evaluation across a range of natural language processing tasks demonstrates that Self-Rewarding PPO consistently outperforms traditional SFT methods. The results highlight the effectiveness of our approach in aligning LLMs using demonstration data, particularly in scenarios where high-quality annotated data is scarce.

cs.CL↗

arXiv2025

OpenDerisk: An Industrial Framework for AI-Driven SRE, with Design, Implementation, and Case Studies

The escalating complexity of modern software imposes an unsustainable operational burden on Site Reliability Engineering (SRE) teams, demanding AI-driven automation that can emulate expert diagnostic reasoning. Existing solutions, from traditional AI methods to general-purpose multi-agent systems, fall short: they either lack deep causal reasoning or are not tailored for the specialized, investigative workflows unique to SRE. To address this gap, we present OpenDerisk, a specialized, open-source multi-agent framework architected for SRE. OpenDerisk integrates a diagnostic-native collaboration model, a pluggable reasoning engine, a knowledge engine, and a standardized protocol (MCP) to enable specialist agents to collectively solve complex, multi-domain problems. Our comprehensive evaluation demonstrates that OpenDerisk significantly outperforms state-of-the-art baselines in both accuracy and efficiency. This effectiveness is validated by its large-scale production deployment at Ant Group, where it serves over 3,000 daily users across diverse scenarios, confirming its industrial-grade scalability and practical impact. OpenDerisk is open source and available at https://github.com/derisk-ai/OpenDerisk/

cs.SE↗