Chinnadhurai Sankar

Personal Page: https://chinnadhurai.github.io/

I am a fourth-year Ph.D Candidate at the University of Montreal,  MILA lab advised by  Prof. Yoshua Bengio .

[Research] I am broadly interested in Deep Learning, NLP, Open-domain & Task-Oriented Dialog Systems, Recurrent Architectures, Deep RL, learning efficient text representations.

[Education] I earned my bachelor's degree at  IIT Madras , where I majored in Electrical Engineering and Physics and my master's degree in ECE at  Purdue University .

[Work] In the past, I have been fortunate enough to have interned at Google Brain/AI(2019,2018,2017), Microsoft Research(2018),  Twitter Cortex(2016). In my previous life, I worked at Qualcomm, San Diego for two years (2013-2015) working on LTE and LTE-advanced features.

[Recent 1st Author Publications]

Deep Reinforcing Learning For Modeling Chit-Chat Dialog With Discrete Attributes.
[SIGDIAL 2019, oral,  Best Paper Award  ] [ arxiv ]

TaskMaster-1 Dialog Corpus: Toward a Realistic and Diverse Dataset .
[EMNLP 2019] [ Google AI post blog ] [ arxiv ] [ data ] [ Media Coverage ]

Do Neural Dialog Systems Use the Conversation History Effectively? An Empirical Study .
[ACL 2019, oral,  Best Paper Award  nomination] [ arxiv ] [ code ]

Transferable Neural Projection Representations .
[NAACL 2019] [ arxiv ]

Towards Non-saturating Recurrent Units for Modeling Long-term Dependencies .
[AAAI 2019, spotlight] [ arxiv ]