arXiv subjects

Zhou Yang

Publications and source records attributed to Zhou Yang.

At least 37 records · Page 2Linked to original sources

arXiv2026

PrivCode++: Latent-Conditioned Differentially Private Code Generation for Comprehensive Guarantees

Large language models fine-tuned on instruction-code pairs may memorize and subsequently leak sensitive training data. Existing differentially private (DP) code generation methods primarily protect code snippets while assuming prompts are public, which fails in realistic scenarios where prompts may also contain sensitive information. When prompts cannot be explicitly learned or used during generation, code synthesis suffers from severe utility degradation as well as reduced diversity and fidelity. To address these challenges, we propose PrivCode-Plus, the first work to explore DP code generation where both prompts and code snippets are considered sensitive in LLM fine-tuning. PrivCode-Plus introduces a two-stage DP framework with a Privacy-Free Latent Conditioning module, enabling effective DP fine-tuning and data synthesis without direct access to sensitive prompts or code. Extensive experiments show that PrivCode-Plus achieves substantially higher utility than baselines, remains competitive with the method with relaxing privacy assumptions, and provides stronger privacy guarantees.

cs.CR↗

arXiv2026

Beyond the Mouth: Upper-Face Affective Cues in Audiovisual Sentence Recognition under Acoustic Uncertainty

Face-to-face speech comprehension is inherently multimodal, integrating acoustic signals with visible articulation, facial expression, head motion, and other socially relevant cues. While audiovisual speech systems typically focus on the mouth region as the primary visual source of linguistic information, affective facial expressions are often treated separately as emotion-recognition targets. This paper investigates whether upper-face affective information contributes to audiovisual sentence recognition beyond audio and mouth-region cues, particularly under acoustic degradation. Using the CREMA-D audiovisual emotional speech corpus, we train feature-based sentence classifiers under four cue conditions: audio only (A), audio plus mouth/lower-face features (A+M), audio plus upper-face features (A+U), and audio plus both mouth and upper-face features (A+M+U). Models are evaluated on clean audio and pink-noise conditions at +10 dB, +5 dB, and 0 dB SNR using actor-independent splits. Results show that mouth/lower-face features provide substantial robustness benefits under degraded audio. At 0 dB SNR, A+M improves accuracy over A by 0.0794, with an actor-bootstrap 95% confidence interval of [0.0296, 0.1298]. Upper-face affective cues exhibit a more nuanced effect. Although the direct accuracy gain of A+M+U over A+M is small, full-face models consistently improve calibration across SNR levels and outperform shuffled upper-face controls under noisy conditions. These findings suggest that affective facial information may support multimodal robustness and confidence estimation under acoustic uncertainty without directly encoding lexical content. More broadly, the study highlights the potential role of socially expressive facial cues in human-centered audiovisual interaction systems.

cs.SD↗

arXiv2026

Revis: Sparse Latent Steering to Mitigate Object Hallucination in Large Vision-Language Models

Despite the advanced capabilities of Large Vision-Language Models (LVLMs), they frequently suffer from object hallucination. One reason is that visual features and pretrained textual representations often become intertwined in the deeper network layers. To address this, we propose REVIS, a training-free framework designed to explicitly re-activate this suppressed visual information. Rooted in latent space geometry, REVIS extracts the pure visual information vector via orthogonal projection and employs a calibrated strategy to perform sparse intervention only at the precise depth where suppression occurs. This surgical approach effectively restores visual information with minimal computational cost. Empirical evaluations on standard benchmarks demonstrate that REVIS reduces object hallucination rates by approximately 19% compared to state-of-the-art baselines, while preserving general reasoning capabilities.

cs.AI↗

arXiv2026

Watermarking LLM Agent Trajectories

LLM agents rely heavily on high-quality trajectory data to guide their problem-solving behaviors, yet producing such data requires substantial task design, high-capacity model generation, and manual filtering. Despite the high cost of creating these datasets, existing literature has overlooked copyright protection for LLM agent trajectories. This gap leaves creators vulnerable to data theft and makes it difficult to trace misuse or enforce ownership rights. This paper introduces ActHook, the first watermarking method tailored for agent trajectory datasets. Inspired by hook mechanisms in software engineering, ActHook embeds hook actions that are activated by a secret input key and do not alter the original task outcome. Like software execution, LLM agents operate sequentially, allowing hook actions to be inserted at decision points without disrupting task flow. When the activation key is present, an LLM agent trained on watermarked trajectories can produce these hook actions at a significantly higher rate, enabling reliable black-box detection. Experiments on mathematical reasoning, web searching, and software engineering agents show that ActHook achieves an average detection AUC of 94.3 on Qwen-2.5-Coder-7B while incurring negligible performance degradation.

cs.CR↗

arXiv2026

Bias in Large Language Models: Origin, Evaluation, and Mitigation

Large Language Models (LLMs) have revolutionized natural language processing, but their susceptibility to biases poses significant challenges. This comprehensive review examines the landscape of bias in LLMs, from its origins to current mitigation strategies. We categorize biases as intrinsic and extrinsic, analyzing their manifestations in various NLP tasks. The review critically assesses a range of bias evaluation methods, including data-level, model-level, and output-level approaches, providing researchers with a robust toolkit for bias detection. We further explore mitigation strategies, categorizing them into pre-model, intra-model, and post-model techniques, highlighting their effectiveness and limitations. Ethical and legal implications of biased LLMs are discussed, emphasizing potential harms in real-world applications such as healthcare and criminal justice. By synthesizing current knowledge on bias in LLMs, this review contributes to the ongoing effort to develop fair and responsible AI systems. Our work serves as a comprehensive resource for researchers and practitioners working towards understanding, evaluating, and mitigating bias in LLMs, fostering the development of more equitable AI technologies.

cs.CL↗

arXiv2026

Industry Practitioners Perspectives on AI Model Quality: Perceptions, Challenges, and Solutions

Artificial Intelligence (AI) is now used across nearly every industry, making AI model quality essential for building reliable and trustworthy systems. Historically, correctness has been the main focus, but industry AI models must also satisfy many other important quality attributes. To understand how these attributes are perceived, the challenges they create, and the solutions used in practice, we identify nine key quality attributes and interview 15 AI practitioners from diverse backgrounds. The interviews show that practitioners prioritize attributes differently depending on context. For example, efficiency can matter more than correctness in real-time applications, while scalability and deployability are no longer seen as primary concerns. Data imbalance emerges as a major obstacle to maintaining model correctness and robustness, and practitioners commonly use mitigation strategies such as active learning. We validate our main findings with a survey of 50 practitioners, which shows that most of the findings are widely recognized. These results can help researchers focus on the attributes practitioners value most and avoid improving one attribute at the expense of others that are considered more critical.

cs.SE↗

arXiv2026

LLMs Are Not a Silver Bullet: A Case Study on Software Fairness

Fairness is a critical requirement for human-related, high-stakes software systems, motivating extensive research on bias mitigation. Prior work has largely focused on tabular data settings using traditional Machine Learning (ML) methods. With the rapid rise of Large Language Models (LLMs), recent studies have begun to explore their use for bias mitigation in the same setting. However, it remains unclear whether LLM-based methods offer advantages over traditional ML methods, leaving software engineers without clear guidance for practical adoption. To address this gap, we present a large-scale study comparing state-of-the-art ML- and LLM-based bias mitigation methods. We find that ML-based methods consistently outperform LLM-based methods in both fairness and predictive performance, with even strong LLMs failing to surpass established ML baselines. To understand why prior LLM-based studies report favorable results, we analyze their evaluation settings and show that these gains are largely driven by artificially balanced test data rather than realistic imbalanced distributions. We further observe that existing LLM-based methods primarily rely on in-context learning and thus fail to leverage all available training data. Motivated by this, we explore supervised fine-tuning on the full training set and find that, while it achieves competitive results, its advantages over traditional ML methods remain limited. These findings suggest that LLMs are not a silver bullet for software fairness.

cs.SE↗

arXiv2026

Can LLMs Deobfuscate Binary Code? A Systematic Analysis of Large Language Models into Pseudocode Deobfuscation

Deobfuscating binary code remains a fundamental challenge in reverse engineering, as obfuscation is widely used to hinder analysis and conceal program logic. Although large language models (LLMs) have shown promise in recovering semantics from obfuscated binaries, a systematic evaluation of their effectiveness is still lacking. In this work, we present BinDeObfBench, the first comprehensive benchmark for assessing LLM-based binary deobfuscation across diverse transformations spanning pre-compilation, compile-time, and post-compilation stages. Our evaluation shows that deobfuscation performance depends more on reasoning capability and domain expertise than on model scale, and that task-specific supervised fine-tuning consistently outperforms broad domain pre-training. Reasoning models can maintain robustness under severe obfuscation, generalize across different instruction set architectures (ISAs) and optimization levels. In-context learning benefits standard models but yields limited gains for reasoning models. Overall, our study highlights the importance of task-specific fine-tuning and reasoning-driven strategies, and positions BinDeObfBench as a basis for future work in binary deobfuscation.

cs.SE↗

arXiv2026

Decoding coherent errors in toric codes on honeycomb and square lattices: duality to Majorana monitored dynamics and symmetry classes

Topological stabilizer codes, such as the toric and surface codes, are leading candidates for fault-tolerant quantum computation. While their decodability under stochastic noise has been extensively studied, the effects of coherent errors, which involve quantum interference, remain less explored. In this work, we study the decodability of toric codes on honeycomb and square lattices subject to $X$- and $Z$-type coherent errors generated by the $X$- and $Z$-rotations on each qubit. We establish a duality between these decoding problems and 1+1D monitored dynamics of non-interacting Majorana fermions. This duality shows that the Altland-Zirnbauer symmetry class of the dual Majorana dynamics governs the universal structure of the decodability phase diagram. We show that the honeycomb-lattice toric code (hTC) with $X$-type error is dual to class-DIII dynamics, while the hTC with $Z$-type error and the square-lattice toric code (sTC) with both error types are dual to class-D dynamics. The key distinction arises from time-reversal symmetry. In class DIII, the generic transition out of the decodable phase is dual to a measurement-induced transition between dynamical phases with area-law and logarithmic entanglement scaling. In contrast, in class D, the generic decodability transition corresponds to a transition between two topologically distinct area-law phases. To explore these transitions in microscopic models, we consider hTC and sTC with $X$-type errors as representatives and introduce a minimal two-parameter coherent error model with spatially varying rotation angles. Using analytical and numerical methods, we map out the decodability phase diagrams and characterize the universal behavior of the transitions. We find that the decodability of sTC is more vulnerable to spatially varying coherent errors than uniform ones.

cond-mat.stat-mech↗

arXiv2026

AgentSZZ: Teaching the LLM Agent to Play Detective with Bug-Inducing Commits

The SZZ algorithm is the dominant technique for identifying bug-inducing commits and underpins many software engineering tasks, such as defect prediction and vulnerability analysis. Despite numerous variants, including recent LLM-based approaches, performance remains limited on developer-annotated datasets (e.g., recall of 0.552 on the Linux kernel). A key limitation is the reliance on git blame, which traces line-level changes within the same file, failing in common scenarios such as ghost and cross-file cases-making nearly one-quarter of bug-inducing commits inherently untraceable. Moreover, current approaches follow fixed pipelines that restrict iterative reasoning and exploration, unlike developers who investigate bugs through an interactive, multi-tool process. To address these challenges, we propose AgentSZZ, an agent-based framework that leverages LLM-driven agents to explore repositories and identify bug-inducing commits. Unlike prior methods, AgentSZZ integrates task-specific tools, domain knowledge, and a ReAct-style loop to enable adaptive and causal tracing of bugs. A structured compression module further improves efficiency by reducing redundant context while preserving key evidence. Extensive experiments on three widely used datasets show that AgentSZZ consistently outperforms state-of-the-art SZZ algorithms across all settings, achieving F1-score gains of up to 27.2% over prior LLM-based approaches. The improvements are especially pronounced in challenging scenarios such as cross-file and ghost commits, with recall gains of up to 300% and 60%, respectively. Ablation studies show that task-specific tools and domain knowledge are critical, while compression tool outputs reduce token consumption by over 30% with negligible impact. The replication package is available.

cs.SE↗

arXiv2026

AEGIS: From Clues to Verdicts -- Graph-Guided Deep Vulnerability Reasoning via Dialectics and Meta-Auditing

Large Language Models (LLMs) are increasingly adopted for vulnerability detection, yet their reasoning remains fundamentally unsound. We identify a root cause shared by both major mitigation paradigms (agent-based debate and retrieval augmentation): reasoning in an ungrounded deliberative space that lacks a bounded, hypothesis-specific evidence base. Without such grounding, agents fabricate cross-function dependencies, and retrieval heuristics supply generic knowledge decoupled from the repository's data-flow topology. Consequently, the resulting conclusions are driven by rhetorical persuasiveness rather than verifiable facts. To ground this deliberation, we present AEGIS, a novel multi-agent framework that shifts detection from ungrounded speculation to forensic verification over a closed factual substrate. Guided by a "From Clue to Verdict" philosophy, AEGIS first identifies suspicious code anomalies (clues), then dynamically reconstructs per-variable dependency chains for each clue via on-demand slicing over a repository-level Code Property Graph. Within this closed evidence boundary, a Verifier Agent constructs competing dialectical arguments for and against exploitability, while an independent Audit Agent scrutinizes every claim against the trace, exercising veto power to prevent hallucinated verdicts. Evaluation on the rigorous PrimeVul dataset demonstrates that AEGIS establishes a new state-of-the-art, achieving 122 Pair-wise Correct Predictions. To our knowledge, this is the first approach to surpass 100 on this benchmark. It reduces the false positive rate by up to 54.40% compared to leading baselines, at an average cost of $0.09 per sample without any task-specific training.

cs.SE↗

arXiv2026

Iterative Semantic Reasoning from Individual to Group Interests for Generative Recommendation with LLMs

Recommendation systems aim to learn user interests from historical behaviors and deliver relevant items. Recent methods leverage large language models (LLMs) to construct and integrate semantic representations of users and items for capturing user interests. However, user behavior theories suggest that truly understanding user interests requires not only semantic integration but also semantic reasoning from explicit individual interests to implicit group interests. To this end, we propose an Iterative Semantic Reasoning Framework (ISRF) for generative recommendation. ISRF leverages LLMs to bridge explicit individual interests and implicit group interests in three steps. First, we perform multi-step bidirectional reasoning over item attributes to infer semantic item features and build a semantic interaction graph capturing users' explicit interests. Second, we generate semantic user features based on the semantic item features and construct a similarity-based user graph to infer the implicit interests of similar user groups. Third, we adopt an iterative batch optimization strategy, where individual explicit interests directly guide the refinement of group implicit interests, while group implicit interests indirectly enhance individual modeling. This iterative process ensures consistent and progressive interest reasoning, enabling more accurate and comprehensive user interest learning. Extensive experiments on the Sports, Beauty, and Toys datasets demonstrate that ISRF outperforms state-of-the-art baselines. The code is available at https://github.com/htired/ISRF.

cs.IR↗

arXiv2026

Automated TEE Adaptation with LLMs: Identifying, Transforming, and Porting Sensitive Functions in Programs

Trusted Execution Environments (TEEs) isolate a special space within a device memory that is not accessible to the normal world (also known as the untrusted environment), even when the device is compromised. Therefore, developers can utilize TEEs to provide robust security guarantees for their programs, protecting sensitive operations, such as encrypted data storage, fingerprint verification, and remote attestation, from software-based attacks. Despite the robust protections offered by TEEs, adapting existing programs to leverage such security guarantees is challenging, often requiring extensive domain knowledge and manual intervention, which makes TEEs less accessible to developers. This motivates us to design AUTOTEE, the first Large Language Model (LLM) enabled approach that can automatically identify, transform, and port functions containing sensitive operations into TEEs with minimal developer intervention. By manually reviewing 68 repositories, we constructed a benchmark dataset consisting of 385 sensitive functions eligible for transformation, on which AUTOTEE achieves a F1 score of 0.94 on Java and 0.87 on Python. AUTOTEE effectively transforms these sensitive functions into TEE-compatible versions, achieving success rates of 91.8% and 84.3% for Java and Python, respectively, when using GPT-4o.

cs.CR↗

arXiv2026

Hidden Licensing Risks in the LLMware Ecosystem

Large Language Models (LLMs) are increasingly integrated into software systems, giving rise to a new class of systems referred to as LLMware. Beyond traditional source-code components, LLMware embeds or interacts with LLMs that depend on other models and datasets, forming complex supply chains across open-source software (OSS), models, and datasets. However, licensing issues emerging from these intertwined dependencies remain largely unexplored. Leveraging GitHub and Hugging Face, we curate a large-scale dataset capturing LLMware supply chains, including 12,180 OSS repositories, 3,988 LLMs, and 708 datasets. Our analysis reveals that license distributions in LLMware differ substantially from traditional OSS ecosystems. We further examine license-related discussions and find that license selection and maintenance are the dominant concerns, accounting for 84% of cases. To understand incompatibility risks, we analyze license conflicts along supply chains and evaluate state-of-the-art detection approaches, which achieve only 58% and 76% F1 scores in this setting. Motivated by these limitations, we propose LiAgent, an LLM-based agent framework for ecosystem-level license compatibility analysis. LiAgent achieves an F1 score of 87%, improving performance by 14 percentage points over prior methods. We reported 60 incompatibility issues detected by LiAgent, 11 of which have been confirmed by developers. Notably, two conflicted LLMs have over 107 million and 5 million downloads on Hugging Face, respectively, indicating potentially widespread downstream impact. We conclude with implications and recommendations to support the sustainable growth of the LLMware ecosystem.

cs.SE↗

arXiv2026

Light Alignment Improves LLM Safety via Model Self-Reflection with a Single Neuron

The safety of large language models (LLMs) has increasingly emerged as a fundamental aspect of their development. Existing safety alignment for LLMs is predominantly achieved through post-training methods, which are computationally expensive and often fail to generalize well across different models. A small number of lightweight alignment approaches either rely heavily on prior-computed safety injections or depend excessively on the model's own capabilities, resulting in limited generalization and degraded efficiency and usability during generation. In this work, we propose a safety-aware decoding method that requires only low-cost training of an expert model and employs a single neuron as a gating mechanism. By effectively balancing the model's intrinsic capabilities with external guidance, our approach simultaneously preserves utility and enhances output safety. It demonstrates clear advantages in training overhead and generalization across model scales, offering a new perspective on lightweight alignment for the safe and practical deployment of large language models. Code: https://github.com/Beijing-AISI/NGSD.

cs.AI↗

arXiv2025

Discovering Transmission Dynamics of COVID-19 in China

A comprehensive retrospective analysis of public health interventions, such as large scale testing, quarantining, and contact tracing, can help identify mechanisms most effective in mitigating COVID-19. We investigate China based SARS-CoV-2 transmission patterns (e.g., infection type and likely transmission source) using publicly released tracking data. We collect case reports from local health commissions, the Chinese CDC, and official local government social media, then apply NLP and manual curation to construct transmission/tracking chains. We further analyze tracking data together with Wuhan population mobility data to quantify and visualize temporal and spatial spread dynamics. Results indicate substantial regional differences, with larger cities showing more infections, likely driven by social activities. Most symptomatic individuals (79\%) were hospitalized within 5 days of symptom onset, and those with confirmed-case contact sought admission in under 5 days. Infection sources also shifted over time: early cases were largely linked to travel to (or contact with travelers from) Hubei Province, while later transmission was increasingly associated with social activities.

cs.LG↗

arXiv2025

Token Sugar: Making Source Code Sweeter for LLMs through Token-Efficient Shorthand

Large language models (LLMs) have shown exceptional performance in code generation and understanding tasks, yet their high computational costs hinder broader adoption. One important factor is the inherent verbosity of programming languages, such as unnecessary formatting elements and lengthy boilerplate code. This leads to inflated token counts in both input and generated outputs, which increases inference costs and slows down the generation process. Prior work improves this through simplifying programming language grammar, reducing token usage across both code understanding and generation tasks. However, it is confined to syntactic transformations, leaving significant opportunities for token reduction unrealized at the semantic level. In this work, we propose Token Sugar, a concept that replaces frequent and verbose code patterns with reversible, token-efficient shorthand in the source code. To realize this concept in practice, we designed a systematic solution that mines high-frequency, token-heavy patterns from a code corpus, maps each to a unique shorthand, and integrates them into LLM pretraining via code transformation. With this solution, we obtain 799 (code pattern, shorthand) pairs, which can reduce up to 15.1% token count in the source code and is complementary to existing syntax-focused methods. We further trained three widely used LLMs on Token Sugar-augmented data. Experimental results show that these models not only achieve significant token savings (up to 11.2% reduction) during generation but also maintain near-identical Pass@1 scores compared to baselines trained on unprocessed code.

cs.SE↗

arXiv2025

MASE: Interpretable NLP Models via Model-Agnostic Saliency Estimation

Deep neural networks (DNNs) have made significant strides in Natural Language Processing (NLP), yet their interpretability remains elusive, particularly when evaluating their intricate decision-making processes. Traditional methods often rely on post-hoc interpretations, such as saliency maps or feature visualization, which might not be directly applicable to the discrete nature of word data in NLP. Addressing this, we introduce the Model-agnostic Saliency Estimation (MASE) framework. MASE offers local explanations for text-based predictive models without necessitating in-depth knowledge of a model's internal architecture. By leveraging Normalized Linear Gaussian Perturbations (NLGP) on the embedding layer instead of raw word inputs, MASE efficiently estimates input saliency. Our results indicate MASE's superiority over other model-agnostic interpretation methods, especially in terms of Delta Accuracy, positioning it as a promising tool for elucidating the operations of text-based models in NLP.

cs.CL↗