Location : Auditing the Black Box: Tracing Unauthori...
题目：Auditing the Black Box: Tracing Unauthorized Data Use in RAG Systems
时间：2025年6月11日（周三）9:40-10:20
地点：中科院数学院南楼N212
报告人：Jiamou Liu（The University of Auckland）
报告摘要：
As large language models (LLMs) are increasingly deployed in Retrieval-Augmented Generation (RAG) systems, the risk of unauthorized data usage—especially personal or sensitive information—has emerged as a significant concern. In this talk, I will present S-RAG, a novel auditing framework designed to detect and attribute the misuse of private data in RAG pipelines. By simulating structured queries and analyzing the LLM’s responses through a set of precision-aligned detectors, S-RAG enables the identification of documents that contribute to unauthorized generations. Our method balances detection accuracy with user-level privacy constraints, and is robust against both white-box and black-box generation models. Building on this foundation, I will introduce a follow-up line of work that extends auditing capabilities to Graph RAG systems, where external structured knowledge (e.g., knowledge graphs) is integrated into LLM generation. We explore how graph structure impacts attribution, and propose new techniques for auditing graph-based retrieval traces.
报告人简介：
Jiamou Liu is an Associate Professor at the School of Computer Science, The University of Auckland, New Zealand. His research spans across several directions in artificial intelligence, in particular, multi-agent systems, algorithmic mechanism design, data marketplaces, and natural language processing. He has recently made contributions to fairness-aware auction mechanisms and privacy-preserving data trading protocols. Dr. Liu has authored over 140 peer-reviewed publications in top-tier venues such as AAAI, AAMAS, IJCAI, WWW, SIGIR, and ICML. He is a recipient of multiple research grants, including Marsden Fund projects and national funding from New Zealand.
CAS,Research Group of Meta-Synthesis and Knowledge Science 京ICP备05002806号-6 文保网安备案号 1101080081 邮箱: mcs@iss.ac.cn
电话：+86 10 82541801