Satwik Kottur bio photo

Satwik Kottur

Research Scientist
Facebook AI
Menlo Park, CA

Email Facebook LinkedIn Github

Welcome

Hi!

I am a research scientist at Facebook AI working in the fields of computer vision, natural language, and machine learning. Prior to joining Facebook, I got my PhD from the Department of ECE, Carnegie Mellon University, advised by Prof. José Moura.

My interests are in the intersection of natural language and computer vision, and broadly, to solve multimodal AI problems using deep learning techniques. I closely collaborate with Prof. Devi Parikh and Prof. Dhruv Batra from Georgia Tech on tasks that require multimodal reasoning (vision and language), for example, visual dialog (VisDial).

I completed my undergraduate from Department of Electrical Engineering (EE), Indian Institute of Technology, Bombay, in 2014 along with a minor in Computer Science and Engineering (CSE).


News

[Nov 2020] Awarded the prestigious A. G. Milnes Award, 2020 by ECE, Carnegie Mellon for my thesis
[Oct 2020] Our paper, Situated and Interactive Multimodal Conversations, has been accepted to COLING, 2020 as an oral presentation
[July 2019] Joined as a research scientist at Facebook AI, Menlo Park (Webpage)
[Mar 2019] Our paper, CLEVR-Dialog: A Diagnostic Dialog Dataset, has been accepted to NAACL, 2019 as an oral presentation
[July 2018] Our paper on Visual Coreference Resolution in Visual Dialog has been accepted to ECCV, 2018
[Jan 2018] Interning at Google Brain this summer, with Gal Chechik and Samy Bengio
[Dec 2017] Awarded the inaugural Snap Inc. Research Fellowship, 2017
[Dec 2017] Received a Best Reviewer Award at NIPS 2017
[Sept 2017] Our paper on Deepsets has been accepted as an oral to NIPS, 2017
[Aug 2017] Our paper on Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog won the best short paper award to EMNLP 2017
[Aug 2017] Our paper on Learning Cooperative Visual Dialog Agents with Deep Reinforcement Learning has been accepted as an oral to ICCV, 2017
[July 2017] Our paper on Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog has been accepted as an oral to EMNLP, 2017
[June 2017] Checkout our latest paper, Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog, on Arxiv. Github code available here
[May 2017] Our paper on Canopy got accepted to ICML, 2017
[Apr 2017] Our paper on Exploring Personalized Neural Conversational Models got accepted at IJCAI, 2017
[Apr 2017] Find our Github code for Visual Dialog here
[Mar 2017] Checkout our latest paper, Learning Cooperative Visual Dialog Agents with Deep Reinforcement Learning, on arXiv
[Mar 2017] Checkout our latest paper, Deepsets, on arXiv
[Mar 2017] Our paper on Visual Dialog got accepted at CVPR, 2017 as a Spotlight
[Mar 2017] Interning at Facebook AI Research (FAIR) this summer, with Marcus Rohrbach
[Nov 2016] Checkout our latest paper, Visual Dialog, on arXiv
[Aug 2016] Serving as Vice-President of ECE Graduate Organization, CMU
[May 2016] Serving as a reviewer for NIPS 2016
[Mar 2016] Our paper on Visual Word2Vec got accepted to CVPR, 2016
[Dec 2015] Interning at Snapchat this summer
[Nov 2015] Checkout our latest paper, Visual Word2Vec, on arXiv
[Oct 2015] Checkout our paper, Comparing Gibbs, EM and SEM for MAP Inference in Mixture Models, which got accepted in NIPS: OPT workshop, 2015

Affiliations

Carnegie Mellon University
2014-2019
IIT Bombay
2010-2014
Viterbi School, USC
Summer 2013
Snapchat Research
Summer 2016
Facebook AI Research
Summer 2017
Google Brain
Summer 2018