Sai Rajeswar

I am a Senior Research Scientist at ServiceNow Research, Montreal team. I completed my Ph.D. in Computer Science at Mila in Université de Montréal, under the supervision of Prof. Aaron Courville. During my Ph.D., I spent time at Google DeepMind as a Research Scientist intern and at ElementAI as a visiting researcher.

Earlier, I graduated from IIT Delhi, with masters in Computer Science. I worked as a Research Engineer at Xerox Research Centre before joining my Ph.D. Previously, I also graduated with an MSc in Mathematics.

IMG-4702 (2)


I am primarily interested in studying how structure helps improve generalization in perception, control and planning.  Recently, I have been focusing on effective pre-training and adaptation strategies for large multimodal models and large-action models. Please see some of the works below for specifics.

Key Topics: world models, reasoning & embodied agents, multimodal AI.


 The best way to reach me is via email at sai dot mudumba at servicenow dot com


Efficient Dynamics Modeling in Interactive Environments with Koopman Theory
Arnab Mondal, Siba Smarak, Sai Rajeswar, Kaleem Siddiqi, Ravanbakhsh
ICLR 2024 

Equivariant Adaptation of Large Pre-trained Models
Arnab Kumar Mondal, Siba Smarak Panigrahi, Sékou-Oumar Kaba, Sai Rajeswar, Siamak Ravanbakhsh.
NeurIPS 2023

Mastering the Unsupervised Reinforcement Learning Benchmark from Pixels
Sai Rajeswar*, Pietro Mazzaglia*, Tim Verbelen, Alexandre Piche, Bart Dhoedt, Aaron Courville, Alexandre Lacoste.
ICML 2023 (oral)
paper / page

Choreographer: Learning and Adapting Skills in Imagination 
Pietro Mazzaglia, Tim Verbelen,  Bart Dhoedt,  Alexandre Lacoste, Sai Rajeswar
ICLR 2023 (top 25% – spotlight)(oral)
paper / code

Hyperbolic Deep Reinforcement Learning for Continuous Control 
Omar Salemohamed, Edoardo Cetin, Sai Rajeswar, Arnab Kumar Mondal
ICLR 2023 (Tiny paper)
paper / code (soon)

Consistency-CAM: Towards Improved Weakly Supervised Semantic Segmentation                           Sai Rajeswar, Issam Lardiji, Pau Rodriguez,  David Vazquez, Aaron Courville                                BMVC, 2022

Unsupervised Model-based Pretraining for Data-Efficient Control from Pixels
Sai Rajeswar*, Pietro Mazzaglia*, Tim Verbelen, Alexandre Piche, Bart Dhoedt, Aaron Courville, Alexandre Lacoste.
DARL workshop, @ICML 2022
paper / code (soon)

Multi-label Iterated Learning for Image Classification with Label Ambiguity                                    Sai Rajeswar*, Pau Rodriguez*, Soumye Singhal, David Vazquez, Aaron Courville                              CVPR, 2022
paper / code


Touch-based Curiosity for Sparse-Reward Tasks
Sai Rajeswar, Cyril Ibrahim, Nitin Surya, Florian Golemo, David Vazquez, Aaron Courville, Pedro O. Pinheiro
CoRL, 2021
paper / code

Pix2shape: Towards unsupervised learning of 3d scenes from images using a view-based representation
Sai Rajeswar, Fahim Mannan, Florian Golemo, Jérôme Parent-Lévesque, David Vazquez, Derek Nowrouzezahrai, Aaron Courville
paper / code

Adversarial computation of optimal transport maps
Jacob Leygonie, Jennifer She, Amjad Almahairi, Sai Rajeswar, Aaron Courville
Preprint, 2019
paper / code

Mutual Information Neural Estimation
Mohamed Ishmael Belghazi, Aristide Baratin, Sai Rajeswar, Sherjil Ozair, Yoshua Bengio, Aaron Courville, R Devon Hjelm
ICML, 2018
paper / code

Augmented cyclegan: Learning many-to-many mappings from unpaired data
Amjad Almahairi, Sai Rajeswar, Alessandro Sordoni, Philip Bachman, Aaron Courville
ICML, 2018
paper / code

Hierarchical adversarially learned inference
Mohamed Ishmael Belghazi*, Sai Rajeswar*, Olivier Mastropietro, Negar Rostamzadeh, Jovana Mitrovic, Aaron Courville
TADGM workshop, @ICML 2018

Towards text generation with adversarially learned neural outlines
Sandeep Subramanian, Sai Rajeswar, Alessandro Sordoni, Adam Trischler, Aaron C Courville, Chris Pal
NeurIPS, 2018

A deep reinforcement learning chatbot
Iulian V Serban, Chinnadhurai Sankar, Mathieu Germain, Saizheng Zhang, Zhouhan Lin, Sandeep Subramanian, Taesup Kim, Michael Pieper, Sarath Chandar, Nan Rosemary Ke, Sai Rajeswar, …, Yoshua Bengio
NeurIPS demo, 2017

Adversarial Generation of Natural Language
Sai Rajeswar*, Sandeep Subramanian*, Francis Dutil, Chris Pal, Aaron Courville
ACL, RepL4NLP, 2017

A hypothesize-and-verify framework for text recognition using deep recurrent neural networks
Anupama Ray, Sai Rajeswar, Santanu Chaudhury
ICDAR, 2015

Cogito, ergo sum