Sarath Chandar

profile img

Sarath Chandar

Assistant Professor
Department of Computer and Software Engineering, École Polytechnique de Montréal.

Core faculty member
Mila, the Quebec Institute of Artificial Intelligence.

Adjunct Professor
Department of Computer Science and Operations Research, Université de Montréal.

Adjunct Professor
Department of Computer Science and Engineering,
Indian Institute of Technology Madras.

Email: fname [DOT] lname [AT] mila [DOT] quebec

Brief Bio:

I did my Ph.D. at University of Montreal, advised by Yoshua Bengioand Hugo Larochelle. I was a recepient of the IBM PhD Fellowshipfor 2018.

Before joining University of Montreal, I was a Research Software Engineer at IBM India Research Lab during August,2014 - May,2015. I obtained my MS by Research at Indian Institute of Technology Madras where I was working with Balaraman Ravindran.


Research Areas:

Recurrent Neural Networks, Lifelong Learning, Reinforcement Learning, Deep Learning, Natural Language Processing, Machine Learning, Artificial Intelligence.

News:

  • NEW(Aug 2020): I will be serving as a Senior Program Committee (SPC) member forAAAI 2021.
  • NEW(Jul 2020): I will be serving as an area chair forICLR 2021.

Recent Publications:

  • NEW MLMLM: Link Prediction with Mean Likelihood Masked Language Model.

    Louis Clouatre, Philippe Trempe, Amal Zouaq, Sarath Chandar.

    In arXiv, 2020.
    [arXiv]

  • NEW How To Evaluate Your Dialogue System: Probe Tasks as an Alternative for Token-level Evaluation Metrics.

    Prasanna Parthasarathi, Joelle Pineau, Sarath Chandar.

    In arXiv, 2020.
    [arXiv],[code]

  • NEW The LoCA Regret: A Consistent Metric to Evaluate Model-Based Behavior in Reinforcement Learning.

    Harm van Seijen, Hadi Nekoei, Evan Racah, Sarath Chandar.

    In arXiv, 2020.
    [arXiv],[code]

  • NEW PatchUp: A Regularization Technique for Convolutional Neural Networks.

    Mojtaba Faramarzi, Mohammad Amini, Akilesh Badrinaaraayanan, Vikas Verma, Sarath Chandar.

    In arXiv, 2020.
    [arXiv],[code]