arXiv subjects

Ronghua Li

Publications and source records attributed to Ronghua Li.

22 records · Page 2Linked to original sources

arXiv2025

"Yes, My LoRD." Guiding Language Model Extraction with Locality Reinforced Distillation

Model extraction attacks (MEAs) on large language models (LLMs) have received increasing attention in recent research. However, existing attack methods typically adapt the extraction strategies originally developed for deep neural networks (DNNs). They neglect the underlying inconsistency between the training tasks of MEA and LLM alignment, leading to suboptimal attack performance. To tackle this issue, we propose Locality Reinforced Distillation (LoRD), a novel model extraction algorithm specifically designed for LLMs. In particular, LoRD employs a newly defined policy-gradient-style training task that utilizes the responses of victim model as the signal to guide the crafting of preference for the local model. Theoretical analyses demonstrate that I) The convergence procedure of LoRD in model extraction is consistent with the alignment procedure of LLMs, and II) LoRD can reduce query complexity while mitigating watermark protection through our exploration-based stealing. Extensive experiments validate the superiority of our method in extracting various state-of-the-art commercial LLMs. Our code is available at: https://github.com/liangzid/LoRD-MEA .

cs.CR↗

arXiv2025

Does Low Rank Adaptation Lead to Lower Robustness against Training-Time Attacks?

Low rank adaptation (LoRA) has emerged as a prominent technique for fine-tuning large language models (LLMs) thanks to its superb efficiency gains over previous methods. While extensive studies have examined the performance and structural properties of LoRA, its behavior upon training-time attacks remain underexplored, posing significant security risks. In this paper, we theoretically investigate the security implications of LoRA's low-rank structure during fine-tuning, in the context of its robustness against data poisoning and backdoor attacks. We propose an analytical framework that models LoRA's training dynamics, employs the neural tangent kernel to simplify the analysis of the training process, and applies information theory to establish connections between LoRA's low rank structure and its vulnerability against training-time attacks. Our analysis indicates that LoRA exhibits better robustness to backdoor attacks than full fine-tuning, while becomes more vulnerable to untargeted data poisoning due to its over-simplified information geometry. Extensive experimental evaluations have corroborated our theoretical findings.

cs.LG↗

arXiv2025

Rethinking Graph Out-Of-Distribution Generalization: A Learnable Random Walk Perspective

Out-Of-Distribution (OOD) generalization has gained increasing attentions for machine learning on graphs, as graph neural networks (GNNs) often exhibit performance degradation under distribution shifts. Existing graph OOD methods tend to follow the basic ideas of invariant risk minimization and structural causal models, interpreting the invariant knowledge across datasets under various distribution shifts as graph topology or graph spectrum. However, these interpretations may be inconsistent with real-world scenarios, as neither invariant topology nor spectrum is assured. In this paper, we advocate the learnable random walk (LRW) perspective as the instantiation of invariant knowledge, and propose LRW-OOD to realize graph OOD generalization learning. Instead of employing fixed probability transition matrix (i.e., degree-normalized adjacency matrix), we parameterize the transition matrix with an LRW-sampler and a path encoder. Furthermore, we propose the kernel density estimation (KDE)-based mutual information (MI) loss to generate random walk sequences that adhere to OOD principles. Extensive experiment demonstrates that our model can effectively enhance graph OOD generalization under various types of distribution shifts and yield a significant accuracy improvement of 3.87% over state-of-the-art graph OOD generalization baselines.

cs.LG↗

arXiv2024

Effective Bug Detection in Graph Database Engines: An LLM-based Approach

Graph database engines play a pivotal role in efficiently storing and managing graph data across various domains, including bioinformatics, knowledge graphs, and recommender systems. Ensuring data accuracy within graph database engines is paramount, as inaccuracies can yield unreliable analytical outcomes. Current bug-detection approaches are confined to specific graph query languages, limiting their applicabilities when handling graph database engines that use various graph query languages across various domains. Moreover, they require extensive prior knowledge to generate queries for detecting bugs. To address these challenges, we introduces DGDB, a novel paradigm harnessing large language models(LLM), such as ChatGPT, for comprehensive bug detection in graph database engines. DGDB leverages ChatGPT to generate high-quality queries for different graph query languages. It subsequently employs differential testing to identify bugs in graph database engines. We applied this paradigm to graph database engines using the Gremlin query language and those using the Cypher query language, generating approximately 4,000 queries each. In the latest versions of Neo4j, Agensgraph, and JanusGraph databases, we detected 2, 5, and 3 wrong-result bugs, respectively.

cs.DB↗