Distributed Hypothesis Testing over Broadcast and Multi-Hop Networks

Han, Csiszar, and Ahlswede, and more recently Rahman and Wagner and Sreekuma and Gunduz, derived fundamental limits on type-II error exponents for distributed binary hypothesis testing when sensors  directly communicate with a single decision center. In this talk, we present similar results for more involved broadcast or multi-hop networks, i.e., for  scenarios where sensors either communicate  with multiple decision centres over a shared link or where they communicate with the  decision center over a network of relays that can also sense and take decisions. The main challenges when dealing with such networks are the competition for network resources and the possibility of improving error exponents by forwarding decisions at intermediate nodes.
We will describe coding schemes for the described networks and present achievable error exponents. We will also prove optimality of the new exponents in special cases of testing against independence and against conditional independence.

The talk is based on joint work with Sadaf Salehkalaibar, Roy Timo, and Ligong Wang.

Bio:
——

Michèle Wigger received the M.Sc. degree in electrical engineering, with distinction, and the Ph.D. degree in electrical engineering both from ETH Zurich in 2003 and 2008, respectively. She was awarded two ETH Medals, one for her Master Thesis and one for her PhD Thesis. Dr. Wigger received a Swiss National Science Foundation Scholarship for Prospective Researchers, and in 2009 was a post-doctoral fellow at the University of California, San Diego, USA. She has held visiting professor appointments at the Technion-Israel Institute of Technology and ETH Zurich. She joined Telecom Paris Tech, Paris France in 2009 where she is currently an Associate Professor. Dr. Wigger is an associate editor for the IEEE Transactions on Information Theory. She is a member of the Board of Governors of the IEEE Information Theory Society. Funding accomplishments include an Emergences Grant from the city of Paris, a Disruptive-ICT grant from Huawei, and an ERC Starting Grant.