June 5, 2017: Invited Speaker Seminar: Inferring User Identity in the Home

Dr. François Schnitzler
Technicolor

Monday, June 5, 2017 at 9:00 am
Room EV001.162

Abstract

The multiplication of connected devices and sensors opens the way to continuous analysis of the domestic context. Knowing which family member does what and where in the house can automate many of the actions of everyday life or improve comfort and recreation through personalization.

This presentation will describe an unsupervised model to identify family members sharing the same accounts or devices. This topic model extends the Latent Dirichlet Allocation using a hidden variable representing the active user and assuming consumption times to be generated by latent time topics. We show that our model is able to learn temporal patterns from the whole set of accounts and infer the active user using both the consumption time and the consumed item.

Time permitting, I will also cover more empirical research to identify family members from data generated by low cost and non-intrusive sensors.

Biography

François Schnitzler is a senior researcher at Technicolor, in the Data Analytics team. He obtained his PhD from the University of Liege in September 2012. His PhD focused on developing probabilistic graphical models for large probability distributions, and in particular ensemble of Markov trees. After his PhD and before joining Technicolor, he was a postdoc fellow and later a senior researcher at the Technion. There, he worked under the supervision of prof. Shie Mannor on the early detection of traffic incidents in Dublin using thousands of sensor data streams. He was involved in time-series modeling and event detection from heterogeneous data and crowdsourcing. He also worked on reinforcement learning.