題目:Heterogeneous Distributed Quickest Change Detection with One-bit Feedback
時間:113年11月15日(星期五13:00- 15:00)
地點:臺北大學三峽校區音律電機資訊大樓電1F02
摘要:
Quickest Change Detection (QCD) problem focuses on the reliable and efficient detection for an abrupt change. Various applications of QCD can be found in distributed learning and IoT in particular. This talk will introduce the Heterogeneous Distributed Quickest Change Detection (HetDQCD) problem with the restriction of one bit feedback. Specifically, the impact of anonymity on the optimal fusion strategy is addressed. We first develop the second order analysis for the performance of anonymous one-shot and voting rule. For the general non-anonymous fusion rule based on the current feedback, we handle the analysis via syndrome-based expression. Eventually, the price of anonymity is shown by comparing both the anonymous and non-anonymous fusion rules.
個人簡介:
Wen-Hsuan Li received the B.S. degree in 2011 and the Ph. D. degree in 2020, all in the department of electrical engineering from National Chiao Tung University, Hsinchu, Taiwan. He is currently a postdoctoral research associate in the institute of communications engineering from National Yang Ming Chiao Tung University, Hsinchu, Taiwan. His research interests include signal processing, cooperative communication, information theory, and sequential change detection.