报告人：张翔   博士后研究员 柏林工业大学

报告题目：

信息论中的安全与隐私：以分层安全聚合和多用户隐私信息检索为例

时间：2024年12月5日10:30-12:00

地点：数学楼2-3会议室

摘要：

Security and privacy have become fundamental aspects in the design of modern information systems. For instance, the rapid growth of artificial intelligence (AI) and its applications has raised significant concerns about data security and user privacy. The training of high-quality AI models depends on massive volumes of data collected from various clients and organizations. However, conventional cryptographic approaches achieve only imperfect security and privacy because they rely on pseudo randomness. Our research focuses on information-theoretic security and privacy, specifically through the lens of zero mutual information, to explore their fundamental impact on various communication and information retrieval systems. In this talk, we will examine two specific applications: (1) secure model aggregation (SecAgg) in federated learning (FL) and (2) private information retrieval (PIR). We will begin by introducing the general problem formulation. Then, we will discuss our recent work on topological SecAgg, which explores SecAgg in a three-layer relay network inspired by hierarchical FL, and cache-aided multiuser PIR, which integrates caching and multiuser considerations into PIR, thereby establishing a connection to coded caching.

报告人简介：

张翔博士于2016年在西安交通大学获得学士学位，并于2023年在美国犹他大学取得硕士和博士学位。目前在柏林工业大学担任博士后研究助理。张博士在IEEE领域内著名学术期刊和国际会议上发表学术论文二十余篇，其中以一作身份发表期刊论文五篇，会议论文若干。张博士长期担任领域内顶级期刊（包括IEEE Transactions on Information Theory，IEEE Transactions on Communication，IEEE Transactions on Wireless Communications等）及会议的审稿人。张博士的研究方向集中于信息论意义上的安全与隐私，及其在多种网络/信息系统中的应用。目前的研究方向包括联邦学习当中的安全聚合（secure aggregation）、缓存辅助的多用户隐私信息检索（cache-aided multiuser private information retrieval）以及WiFi网络当中的高效用户调度和资源分配。

邀请人：魏志强 教授