2018年7月“現代無線通信論壇”（ Workshop on Advances in Wireless Communications 2018）

Prof. Simon Godsill

(University of Cambridge, UK)

Title: Bayesian Learning of latent dynamic network structures and intentionality from noisy sensor data

Abstract: In this talk we motivate and describe methodologies for uncovering dynamic structures, linkages and intentionalities from incomplete and noisy data, potentially arriving from multiple sensors. We illustrate the methods through application to inference about destinations of shipping data, multiple object tracking data where object behaviours and linkages may be evolving over time, high frequency financial data where very noisy and incomplete information must be used to predict and react to changes in the market structure over time, and in User Interfaces for automobiles in which the task is to determine accurately and rapidly the intended icon a user is pointing at on a screen, based on the trajectory of hand motion near to the screen, and in the presence of disturbances from suspension and road surface.

Prof. Sennur Ulukus

(University of Maryland, USA)

Title: Private Information Retrieval

Abstract: Private information retrieval (PIR) is a canonical problem to study the privacy of users as they download content from publicly accessible databases. In PIR, a user (retriever) wishes to download data from one or more databases in such a way that no individual database can tell which data has been retrieved. PIR has originated in the computer science literature, and has recently been revisited by the information theory community. The information-theoretic re formulation of the problem aims at determining the fundamental limits of the PIR problem, i.e., what is the largest number of bits that can be retrieved privately per bit of download, or equivalently, what is the minimum number of downloads needed per bit of private retrieval? This new information-theoretic approach also proposes novel PIR schemes which achieve or approach these fundamental limits. In this talk, I will describe the problem, summarize several break-through results in the history of this problem, and present some of the recent advances in this field.