Akilesh Badrinaaraayanan

I am a Ph.D. student at CMU Tepper School of Business with focus on Information Systems and Marketing. I recently graduated with a M.Sc. by research in Machine Learning from Université de Montréal and Mila - Quebec AI Institute, where I worked on reinforcement learning and lifelong learning supervised by Prof. Aaron Courville and Prof. Sarath Chandar. Before joining Mila, I worked as a MTS at Adobe Systems, India wherein I worked on both fundamental and applied ML problems. Previously, I did my undergrad in computer science from IIT-Hyderabad where I worked with Prof. Vineeth Balasubramanian on some interesting computer vision problems. During my undergrad, I also spent time at Bosch as a research intern working on diabetic retinopathy detection. I have also dabbled with other areas of computer science during my undergrad. I have been extremely fortunate to have some of the best mentors and collaborators in my career thus far.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github

profile photo
Research

I'm interested in machine learning, reinforcement learning, lifelong learning and computer vision.

blind-date Continuous Coordination As a Realistic Scenario for Lifelong Learning
Akilesh Badrinaaraayanan Hadi Nekoei , Aaron Courville, Sarath Chandar,
ICML, 2021 , ICLR NERL workshop, 2021 spotlight
Code

Propose a new benchmark for lifelong RL based on Hanabi as well as introduce lifelong learning as a step to think beyond the centralized-training paradigm in MARL.

blind-date PatchUp: A Regularization Technique for Convolutional Neural Networks
Mojtaba Faramarzi, Mohammad Amini, Akilesh Badrinaaraayanan , Vikas Verma, Sarath Chandar
arXiv, 2020
Code

Novel regularization technique with better empirical performance and robustness than most existing regularization methods across several datasets/architectures.

blind-date Towards Jumpy Planning
Akilesh Badrinaaraayanan , Suriya Singh, Anirudh Goyal, Alexander Neitz, Aaron Courville
ICML MBRL workshop, 2019 spotlight

Developed a model-based planner with a goal-conditioned policy trained with model-free learning, i.e. dynamical models that jump between decision states.

blind-date Attention Based Natural Language Grounding By Navigating Virtual Environment
Akilesh Badrinaaraayanan , Abhishek Sinha, Mausoom Sarkar, Balaji Krishnamurthy.
WACV, 2019, NeurIPS ViGIL workshop, 2017.
Code

Language grounding in 2D/3D environments through multimodal fusion of visual and textual features.