arXiv subjects

Lan Zhang

Publications and source records attributed to Lan Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2025

Importance-Aware Data Selection for Efficient LLM Instruction Tuning

Instruction tuning plays a critical role in enhancing the performance and efficiency of Large Language Models (LLMs). Its success depends not only on the quality of the instruction data but also on the inherent capabilities of the LLM itself. Some studies suggest that even a small amount of high-quality data can achieve instruction fine-tuning results that are on par with, or even exceed, those from using a full-scale dataset. However, rather than focusing solely on calculating data quality scores to evaluate instruction data, there is a growing need to select high-quality data that maximally enhances the performance of instruction tuning for a given LLM. In this paper, we propose the Model Instruction Weakness Value (MIWV) as a novel metric to quantify the importance of instruction data in enhancing model's capabilities. The MIWV metric is derived from the discrepancies in the model's responses when using In-Context Learning (ICL), helping identify the most beneficial data for enhancing instruction tuning performance. Our experimental results demonstrate that selecting only the top 1\% of data based on MIWV can outperform training on the full dataset. Furthermore, this approach extends beyond existing research that focuses on data quality scoring for data selection, offering strong empirical evidence supporting the effectiveness of our proposed method.

cs.CL↗

arXiv2025

Adaptive Semantic Communication for UAV/UGV Cooperative Path Planning

Effective path planning is fundamental to the coordination of unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) systems, particularly in applications such as surveillance, navigation, and emergency response. Combining UAVs' broad field of view with UGVs' ground-level operational capability greatly improve the likelihood of successfully achieving task objectives such as locating victims, monitoring target areas, or navigating hazardous terrain. In complex environments, UAVs need to provide precise environmental perception information for UGVs to optimize their routing policy. However, due to severe interference and non-line-of-sight conditions, wireless communication is often unstable in such complex environments, making it difficult to support timely and accurate path planning for UAV-UGV coordination. To this end, this paper proposes a semantic communication (SemCom) framework to enhance UAV/UGV cooperative path planning under unreliable wireless conditions. Unlike traditional methods that transmit raw data, SemCom transmits only the key information for path planning, reducing transmission volume without sacrificing accuracy. The proposed framework is developed by defining key semantics for path planning and designing a transceiver for meeting the requirements of UAV-UGV cooperative path planning. Simulation results show that, compared to conventional SemCom transceivers, the proposed transceiver significantly reduces data transmission volume while maintaining path planning accuracy, thereby enhancing system collaboration efficiency.

cs.NI↗

arXiv2025

MASA: LLM-Driven Multi-Agent Systems for Autoformalization

Autoformalization serves a crucial role in connecting natural language and formal reasoning. This paper presents MASA, a novel framework for building multi-agent systems for autoformalization driven by Large Language Models (LLMs). MASA leverages collaborative agents to convert natural language statements into their formal representations. The architecture of MASA is designed with a strong emphasis on modularity, flexibility, and extensibility, allowing seamless integration of new agents and tools to adapt to a fast-evolving field. We showcase the effectiveness of MASA through use cases on real-world mathematical definitions and experiments on formal mathematics datasets. This work highlights the potential of multi-agent systems powered by the interaction of LLMs and theorem provers in enhancing the efficiency and reliability of autoformalization, providing valuable insights and support for researchers and practitioners in the field.

cs.CL↗

arXiv2025

What Lurks Within? Concept Auditing for Shared Diffusion Models at Scale

Diffusion models (DMs) have revolutionized text-to-image generation, enabling the creation of highly realistic and customized images from text prompts. With the rise of parameter-efficient fine-tuning (PEFT) techniques, users can now customize powerful pre-trained models using minimal computational resources. However, the widespread sharing of fine-tuned DMs on open platforms raises growing ethical and legal concerns, as these models may inadvertently or deliberately generate sensitive or unauthorized content. Despite increasing regulatory attention on generative AI, there are currently no practical tools for systematically auditing these models before deployment. In this paper, we address the problem of concept auditing: determining whether a fine-tuned DM has learned to generate a specific target concept. Existing approaches typically rely on prompt-based input crafting and output-based image classification but they suffer from critical limitations, including prompt uncertainty, concept drift, and poor scalability. To overcome these challenges, we introduce Prompt-Agnostic Image-Free Auditing (PAIA), a novel, model-centric concept auditing framework. By treating the DM as the object of inspection, PAIA enables direct analysis of internal model behavior, bypassing the need for optimized prompts or generated images. We evaluate PAIA on 320 controlled models trained with curated concept datasets and 771 real-world community models sourced from a public DM sharing platform. Evaluation results show that PAIA achieves over 90% detection accuracy while reducing auditing time by 18 - 40X compared to existing baselines. To our knowledge, PAIA is the first scalable and practical solution for pre-deployment concept auditing of diffusion models, providing a practical foundation for safer and more transparent diffusion model sharing.

cs.LG↗

arXiv2025

Who's Wearing? Ear Canal Biometric Key Extraction for User Authentication on Wireless Earbuds

Ear canal scanning/sensing (ECS) has emerged as a novel biometric authentication method for mobile devices paired with wireless earbuds. Existing studies have demonstrated the uniqueness of ear canals by training and testing machine learning classifiers on ECS data. However, implementing practical ECS-based authentication requires preventing raw biometric data leakage and designing computationally efficient protocols suitable for resource-constrained earbuds. To address these challenges, we propose an ear canal key extraction protocol, \textbf{EarID}. Without relying on classifiers, EarID extracts unique binary keys directly on the earbuds during authentication. These keys further allow the use of privacy-preserving fuzzy commitment scheme that verifies the wearer's key on mobile devices. Our evaluation results demonstrate that EarID achieves a 98.7\% authentication accuracy, comparable to machine learning classifiers. The mobile enrollment time (160~ms) and earbuds processing time (226~ms) are negligible in terms of wearer's experience. Moreover, our approach is robust and attack-resistant, maintaining a false acceptance rate below 1\% across all adversarial scenarios. We believe the proposed EarID offers a practical and secure solution for next-generation wireless earbuds.

cs.CR↗

arXiv2025

Your RAG is Unfair: Exposing Fairness Vulnerabilities in Retrieval-Augmented Generation via Backdoor Attacks

Retrieval-augmented generation (RAG) enhances factual grounding by integrating retrieval mechanisms with generative models but introduces new attack surfaces, particularly through backdoor attacks. While prior research has largely focused on disinformation threats, fairness vulnerabilities remain underexplored. Unlike conventional backdoors that rely on direct trigger-to-target mappings, fairness-driven attacks exploit the interaction between retrieval and generation models, manipulating semantic relationships between target groups and social biases to establish a persistent and covert influence on content generation. This paper introduces BiasRAG, a systematic framework that exposes fairness vulnerabilities in RAG through a two-phase backdoor attack. During the pre-training phase, the query encoder is compromised to align the target group with the intended social bias, ensuring long-term persistence. In the post-deployment phase, adversarial documents are injected into knowledge bases to reinforce the backdoor, subtly influencing retrieved content while remaining undetectable under standard fairness evaluations. Together, BiasRAG ensures precise target alignment over sensitive attributes, stealthy execution, and resilience. Empirical evaluations demonstrate that BiasRAG achieves high attack success rates while preserving contextual relevance and utility, establishing a persistent and evolving threat to fairness in RAG.

cs.IR↗

arXiv2025

DyME: Dynamic Multi-Concept Erasure in Diffusion Models with Bi-Level Orthogonal LoRA Adaptation

Text-to-image diffusion models (DMs) inadvertently reproduce copyrighted styles and protected visual concepts, raising legal and ethical concerns. Concept erasure has emerged as a safeguard, aiming to selectively suppress such concepts through fine-tuning. However, existing methods do not scale to practical settings where providers must erase multiple and possibly conflicting concepts. The core bottleneck is their reliance on static erasure: a single checkpoint is fine-tuned to remove all target concepts, regardless of the actual erasure needs at inference. This rigid design mismatches real-world usage, where requests vary per generation, leading to degraded erasure success and reduced fidelity for non-target content. We propose DyME, an on-demand erasure framework that trains lightweight, concept-specific LoRA adapters and dynamically composes only those needed at inference. This modular design enables flexible multi-concept erasure, but naive composition causes interference among adapters, especially when many or semantically related concepts are suppressed. To overcome this, we introduce bi-level orthogonality constraints at both the feature and parameter levels, disentangling representation shifts and enforcing orthogonal adapter subspaces. We further develop ErasureBench-H, a new hierarchical benchmark with brand-series-character structure, enabling principled evaluation across semantic granularities and erasure set sizes. Experiments on ErasureBench-H and standard datasets (e.g., CIFAR-100, Imagenette) demonstrate that DyME consistently outperforms state-of-the-art baselines, achieving higher multi-concept erasure fidelity with minimal collateral degradation.

cs.CV↗

arXiv2025

Position: Human Factors Reshape Adversarial Analysis in Human-AI Decision-Making Systems

As Artificial Intelligence (AI) increasingly supports human decision-making, its vulnerability to adversarial attacks grows. However, the existing adversarial analysis predominantly focuses on fully autonomous AI systems, where decisions are executed without human intervention. This narrow focus overlooks the complexities of human-AI collaboration, where humans interpret, adjust, and act upon AI-generated decisions. Trust, expectations, and cognitive behaviors influence how humans interact with AI, creating dynamic feedback loops that adversaries can exploit. To strengthen the robustness of AI-assisted decision-making, adversarial analysis must account for the interplay between human factors and attack strategies. This position paper argues that human factors fundamentally reshape adversarial analysis and must be incorporated into evaluating robustness in human-AI decision-making systems. To fully explore human factors in adversarial analysis, we begin by investigating the role of human factors in human-AI collaboration through a comprehensive review. We then introduce a novel robustness analysis framework that (1) examines how human factors affect collaborative decision-making performance, (2) revisits and interprets existing adversarial attack strategies in the context of human-AI interaction, and (3) introduces a new timing-based adversarial attack as a case study, illustrating vulnerabilities emerging from sequential human actions. The experimental results reveal that attack timing uniquely impacts decision outcomes in human-AI collaboration. We hope this analysis inspires future research on adversarial robustness in human-AI systems, fostering interdisciplinary approaches that integrate AI security, human cognition, and decision-making dynamics.

cs.HC↗

arXiv2025

AIP: Subverting Retrieval-Augmented Generation via Adversarial Instructional Prompt

Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by retrieving relevant documents from external sources to improve factual accuracy and verifiability. However, this reliance introduces new attack surfaces within the retrieval pipeline, beyond the LLM itself. While prior RAG attacks have exposed such vulnerabilities, they largely rely on manipulating user queries, which is often infeasible in practice due to fixed or protected user inputs. This narrow focus overlooks a more realistic and stealthy vector: instructional prompts, which are widely reused, publicly shared, and rarely audited. Their implicit trust makes them a compelling target for adversaries to manipulate RAG behavior covertly. We introduce a novel attack for Adversarial Instructional Prompt (AIP) that exploits adversarial instructional prompts to manipulate RAG outputs by subtly altering retrieval behavior. By shifting the attack surface to the instructional prompts, AIP reveals how trusted yet seemingly benign interface components can be weaponized to degrade system integrity. The attack is crafted to achieve three goals: (1) naturalness, to evade user detection; (2) utility, to encourage use of prompts; and (3) robustness, to remain effective across diverse query variations. We propose a diverse query generation strategy that simulates realistic linguistic variation in user queries, enabling the discovery of prompts that generalize across paraphrases and rephrasings. Building on this, a genetic algorithm-based joint optimization is developed to evolve adversarial prompts by balancing attack success, clean-task utility, and stealthiness. Experimental results show that AIP achieves up to 95.23% ASR while preserving benign functionality. These findings uncover a critical and previously overlooked vulnerability in RAG systems, emphasizing the need to reassess the shared instructional prompts.

cs.CV↗

arXiv2025

Autoformalization in the Wild: Assessing LLMs on Real-World Mathematical Definitions

Thanks to their linguistic capabilities, LLMs offer an opportunity to bridge the gap between informal mathematics and formal languages through autoformalization. However, it is still unclear how well LLMs generalize to sophisticated and naturally occurring mathematical statements. To address this gap, we investigate the task of autoformalizing real-world mathematical definitions: a critical component of mathematical discourse. Specifically, we introduce two novel resources for autoformalization, collecting definitions from Wikipedia (Def_Wiki) and arXiv papers (Def_ArXiv). We then systematically evaluate a range of LLMs, analyzing their ability to formalize definitions into Isabelle/HOL. Furthermore, we investigate strategies to enhance LLMs' performance including refinement through external feedback from Proof Assistants, and formal definition grounding, where we augment LLMs' formalizations through relevant contextual elements from formal mathematical libraries. Our findings reveal that definitions present a greater challenge compared to existing benchmarks, such as miniF2F. In particular, we found that LLMs still struggle with self-correction, and aligning with relevant mathematical libraries. At the same time, structured refinement methods and definition grounding strategies yield notable improvements of up to 16% on self-correction capabilities and 43% on the reduction of undefined errors, highlighting promising directions for enhancing LLM-based autoformalization in real-world scenarios.

cs.CL↗

arXiv2025

Regularity and dynamics of weak solutions for one-dimensional compressible Navier-Stokes equations with vacuum

In the spirit of D. Hoff's weak solution theory for the compressible Navier-Stokes equations (CNS) with bounded density, in this paper we establish the global existence and regularity properties of finite-energy weak solutions to an initial boundary value problem of one-dimensional CNS with general initial data and vacuum. The core of our proof is a global in time a priori estimate for one-dimensional CNS that holds for any $H^1$ initial velocity and bounded initial density not necessarily strictly positive: it could be a density patch or a vacuum bubble. We also establish that the velocity and density decay exponentially to equilibrium. As a by-product, we obtain the quantitative dynamics of aforementioned two vacuum states.

math.AP↗

arXiv2025

Uncover 3D Dark Matter Distribution of the Milky Way by an Empirical Triaxial Orbit-Superposition Model: Method Validation

We introduce a novel dynamical model, named empirical triaxial orbit-superposition model, for the Milky Way halo. This model relies on minimal physical assumptions that the system is stationary, meaning the distribution function in 6D phase-space does not change when the stars orbiting in the correct gravitational potential. We validate our method by applying it to mock datasets that mimic the observations of the Milky Way halo from LAMOST + Gaia with stars' 3D position and 3D velocity observed. By removing the stellar disk and substructures, correcting the selection function, we obtain a sample of smooth halo stars considered as stationary and complete. We construct a gravitational potential including a highly flexible triaxial dark matter halo with adaptable parameters. Within each specified gravitational potential, we integrate orbits of these halo stars, and build a model by superposing the orbits together taking the weights of stars derived from the selection function correction. The goodness of the models are evaluated by comparing the density distributions as well as 3D velocity distributions numerically represented in the model to that in the data. The shape and radial density distribution of the underlying dark matter halo can be constrained well simultaneously. We apply it to three mock galaxies with different intrinsic shapes of their dark matter halos and achieved accurate recovery of the 3D dark matter density distributions for all.

astro-ph.GA↗

arXiv2025

Untraceable DeepFakes via Traceable Fingerprint Elimination

Recent advancements in DeepFakes attribution technologies have significantly enhanced forensic capabilities, enabling the extraction of traces left by generative models (GMs) in images, making DeepFakes traceable back to their source GMs. Meanwhile, several attacks have attempted to evade attribution models (AMs) for exploring their limitations, calling for more robust AMs. However, existing attacks fail to eliminate GMs' traces, thus can be mitigated by defensive measures. In this paper, we identify that untraceable DeepFakes can be achieved through a multiplicative attack, which can fundamentally eliminate GMs' traces, thereby evading AMs even enhanced with defensive measures. We design a universal and black-box attack method that trains an adversarial model solely using real data, applicable for various GMs and agnostic to AMs. Experimental results demonstrate the outstanding attack capability and universal applicability of our method, achieving an average attack success rate (ASR) of 97.08\% against 6 advanced AMs on DeepFakes generated by 9 GMs. Even in the presence of defensive mechanisms, our method maintains an ASR exceeding 72.39\%. Our work underscores the potential challenges posed by multiplicative attacks and highlights the need for more robust AMs.

cs.CR↗

arXiv2025

Security Tensors as a Cross-Modal Bridge: Extending Text-Aligned Safety to Vision in LVLM

Large visual-language models (LVLMs) integrate aligned large language models (LLMs) with visual modules to process multimodal inputs. However, the safety mechanisms developed for text-based LLMs do not naturally extend to visual modalities, leaving LVLMs vulnerable to harmful image inputs. To address this cross-modal safety gap, we introduce security tensors - trainable input vectors applied during inference through either the textual or visual modality. These tensors transfer textual safety alignment to visual processing without modifying the model's parameters. They are optimized using a curated dataset containing (i) malicious image-text pairs requiring rejection, (ii) contrastive benign pairs with text structurally similar to malicious queries, with the purpose of being contrastive examples to guide visual reliance, and (iii) general benign samples preserving model functionality. Experimental results demonstrate that both textual and visual security tensors significantly enhance LVLMs' ability to reject diverse harmful visual inputs while maintaining near-identical performance on benign tasks. Further internal analysis towards hidden-layer representations reveals that security tensors successfully activate the language module's textual "safety layers" in visual inputs, thereby effectively extending text-based safety to the visual modality.

cs.CV↗

arXiv2025

IntentVCNet: Bridging Spatio-Temporal Gaps for Intention-Oriented Controllable Video Captioning

Intent-oriented controlled video captioning aims to generate targeted descriptions for specific targets in a video based on customized user intent. Current Large Visual Language Models (LVLMs) have gained strong instruction following and visual comprehension capabilities. Although the LVLMs demonstrated proficiency in spatial and temporal understanding respectively, it was not able to perform fine-grained spatial control in time sequences in direct response to instructions. This substantial spatio-temporal gap complicates efforts to achieve fine-grained intention-oriented control in video. Towards this end, we propose a novel IntentVCNet that unifies the temporal and spatial understanding knowledge inherent in LVLMs to bridge the spatio-temporal gap from both prompting and model perspectives. Specifically, we first propose a prompt combination strategy designed to enable LLM to model the implicit relationship between prompts that characterize user intent and video sequences. We then propose a parameter efficient box adapter that augments the object semantic information in the global visual context so that the visual token has a priori information about the user intent. The final experiment proves that the combination of the two strategies can further enhance the LVLM's ability to model spatial details in video sequences, and facilitate the LVLMs to accurately generate controlled intent-oriented captions. Our proposed method achieved state-of-the-art results in several open source LVLMs and was the runner-up in the IntentVC challenge. Our code is available on https://github.com/thqiu0419/IntentVCNet.

cs.CV↗

arXiv2025

Sample-Efficient Reinforcement Learning Controller for Deep Brain Stimulation in Parkinson's Disease

Deep brain stimulation (DBS) is an established intervention for Parkinson's disease (PD), but conventional open-loop systems lack adaptability, are energy-inefficient due to continuous stimulation, and provide limited personalization to individual neural dynamics. Adaptive DBS (aDBS) offers a closed-loop alternative, using biomarkers such as beta-band oscillations to dynamically modulate stimulation. While reinforcement learning (RL) holds promise for personalized aDBS control, existing methods suffer from high sample complexity, unstable exploration in binary action spaces, and limited deployability on resource-constrained hardware. We propose SEA-DBS, a sample-efficient actor-critic framework that addresses the core challenges of RL-based adaptive neurostimulation. SEA-DBS integrates a predictive reward model to reduce reliance on real-time feedback and employs Gumbel Softmax-based exploration for stable, differentiable policy updates in binary action spaces. Together, these components improve sample efficiency, exploration robustness, and compatibility with resource-constrained neuromodulatory hardware. We evaluate SEA-DBS on a biologically realistic simulation of Parkinsonian basal ganglia activity, demonstrating faster convergence, stronger suppression of pathological beta-band power, and resilience to post-training FP16 quantization. Our results show that SEA-DBS offers a practical and effective RL-based aDBS framework for real-time, resource-constrained neuromodulation.

cs.LG↗

arXiv2025

Efficient and Accurate Image Provenance Analysis: A Scalable Pipeline for Large-scale Images

The rapid proliferation of modified images on social networks that are driven by widely accessible editing tools demands robust forensic tools for digital governance. Image provenance analysis, which filters various query image variants and constructs a directed graph to trace their phylogeny history, has emerged as a critical solution. However, existing methods face two fundamental limitations: First, accuracy issues arise from overlooking heavily modified images due to low similarity while failing to exclude unrelated images and determine modification directions under diverse modification scenarios. Second, scalability bottlenecks stem from pairwise image analysis incurs quadratic complexity, hindering application in large-scale scenarios. This paper presents a scalable end-to-end pipeline for image provenance analysis that achieves high precision with linear complexity. This improves filtering effectiveness through modification relationship tracing, which enables the comprehensive discovery of image variants regardless of their visual similarity to the query. In addition, the proposed pipeline integrates local features matching and compression artifact capturing, enhancing robustness against diverse modifications and enabling accurate analysis of images' relationships. This allows the generation of a directed provenance graph that accurately characterizes the image's phylogeny history. Furthermore, by optimizing similarity calculations and eliminating redundant pairwise analysis during graph construction, the pipeline achieves a linear time complexity, ensuring its scalability for large-scale scenarios. Experiments demonstrate pipeline's superior performance, achieving a 16.7-56.1% accuracy improvement. Notably, it exhibits significant scalability with an average 3.0-second response time on 10 million scale images, which is far shorter than the SOTA approach's 12-minute duration.

cs.MM↗

arXiv2025

Power-Efficient Optimization for Coexisting Semantic and Bit-Based Users in NOMA Networks

Semantic communications, which focus on transmitting the semantic meaning of data, have been proposed as a novel paradigm for achieving efficient and relevant communication. Meanwhile, non-orthogonal multiple access (NOMA) enhances spectral efficiency by allowing multiple users to share the same spectrum. However, semantic communications are unlikely to fully replace conventional bit-level communications in the near future, as the latter remain dominant. Therefore, integrating semantic users into a NOMA network alongside conventional bit-based users becomes a meaningful approach to improve both transmission and spectrum efficiency. Nonetheless, due to the lack of a mathematical model that accurately characterizes the relationship between the performance of semantic transceivers and wireless resource allocation, enhancing performance through resource optimization remains a challenge. Moreover, successive interference cancellation (SIC), a key technique in NOMA, introduces additional complexity in system design and implementation. To address these challenges, this paper first improves the deep semantic communication (DeepSC) transceiver to make it adaptive to varying wireless transmission conditions. Subsequently, a data-driven regression approach is employed to develop a mathematical model that captures the impact of wireless resources on semantic transceiver performance. In parallel, a multi-cluster hybrid NOMA (H-NOMA) framework is proposed, where each cluster consists of one semantic user and one bit-based user, to mitigate the complexity introduced by SIC. A total transmit power minimization problem is then formulated by jointly optimizing the beamforming design, bandwidth allocation, and semantic symbol factor.

eess.SP↗