Paolo Viappiani
homepage:http://www.cs.toronto.edu/~paolo/
search externally:   Google Scholar,   Springer,   CiteSeer,   Microsoft Academic Search,   Scirus ,   DBlife

Description

From 2008 to August 2010 I was a postdoc at the University of Toronto working with Craig Boutilier on topics of utility elicitation, decision making under uncertainty and optimization. I hold a PhD in Artificial Intelligence from EPFL, Lausanne, Switzerland (obtained in 2007). My interests are within the following areas: recommender systems, constraint and planning systems, personal agents, preferences and user modeling, intelligent interfaces, machine learning, adaptive systems.

I am currently investigating preference reasoning and elicitation for automatic tools like personal agents or recommender systems. In our opinion, preference-elicitation should be open-ended allowing a diversity of modalites to acquire user models and target recommendations. Users should be allowed to state preferences in a way that is natural and convenient to them. At the same time, AI systems need to optimize their actions reasoning with a partial model of the user preferences or utility function.


Lecture:

demonstration video
flag Optimal Bayesian Recommendation Sets and Myopically Optimal Choice Query Sets
as coauthor at  Video Journal of Machine Learning Abstracts - Volume 1,
3131 views