Gábor Bartók

I am a Postdoctaral Fellow at ETH Zürich in the Learning & Adaptive Systems Group with Andreas Krause. I finished my PhD at the University of Alberta in 2012, under the supervision of Csaba Szepesvári. I was a member of the Reinforcement Learning and Artificial Intelligence research group.

My research interest is machine learning. My goal is to better understand the theoretical limitations of machine learning. Currently I am focusing on online learning. See my publication list.

News

I accepted a software engineer position at Google Zurich, starting in September 2014.

Our paper with Adish Singla, Ilija Bogunovic, Amin Karbasi and Andreas Krause titled Near-Optimally Teaching the Crowd to Classify, has been accepted at ICML2014.

The University of Alberta nominated my thesis for the WAGS/UMI Innovation in Technology Award. The winners will be announced in March 2014 at the WAGS annual meeting.

Our paper with Navid  Zolghadr, András  György, Csaba  Szepesvári, and Russell  Greiner, titled Online Learning with Costly Features and Labels, has been accepted for poster presentation to NIPS2013.

Our paper with Gergely Neu, An efficient algorithm for learning with semi-bandit feedback, has been accepted to ALT2013.

I won the 2013 Doctoral Dissertation Award of the Canadian Artificial Intelligence Association!

My paper "A near-optimal algorithm for finite partial-monitoring games against adversarial opponents" has been accepted to COLT2013.

My PhD thesis is nominated for the AI Doctoral Dissertation Award of the Canadian Artificial Intelligence Association (CAIAC) by the Department of Computing Science, University of Alberta

I gave a contributed talk at the NIPS2012 Workshop on Information in Perception and Action, titled "The value of information in online learning: A study of partial monitoring problems"

Our paper with Csaba Szepesvári "Partial monitoring with side information" has been accepted to ALT2012.

I defended my PhD thesis on the 22nd of June, 2012. My thesis title is "The role of information in online learning". I will upload the final version soon.

Our ICML2012 submission with Navid and Csaba "An adaptive algorithm for finite stochastic partial monitoring" has been accepted.

From mid-summer 2012, I will work as a postdoctoral fellow with Andreas Krause in the Learning & Adaptive Systems Group at ETH Zürich.