Vedant Shah

I am a Research Master's student at Mila and Université de Montréal where I am supervised by Prof. Yoshua Bengio's and Dr. Anirudh Goyal.

I completed my undergraduate studies in Electronics and Communications Engineering from BITS Pilani, Goa Campus where I was advised by Prof. Ashwin Srinivasan and Prof. Tanmay Verlekar at APPCAIR. I also spent a summer interning with Dr. Gautam Shroff at TCS Research and was selected for Google Summer of Code 2021.

At BITS, I was the Vice President of the Society for Artificial Intelligence and Deep Learning and Teaching Head of the Electronics and Robotics Club, BITS Goa.

Email  /  CV  /  Scholar  /  LinkedIn  /  Twitter  /  Github

Broadly, I am interested in improving the sample efficiency and out of distribution generalization abilities of Reinforcement Learning and Deep Learning agents. To this end, I am interested in looking into approaches based on human cognitive inductive biases, meta learning and neurosymbolic methods. Recently, I have been working on investigating different pretraining approaches for Deep Reinforcement Learning. I have also been working on using probabilistic inference framework of GFlowNet for Drug Discovery.

We propose a nearly compute-free approach for class unlearning in multi-class classification settings.

Investigating the problem of environmental heterogeneity and using behvioural priors to tackle the problems of coordination and heterogeneity in MARL

Using corrupted pretraining and semantic loss for obtaining explainable predictions in terms of human intelligible concepts by composing a symbolic and a neural module together.

We train state of the art deep learning models on financial data using a data augmentation with synthetic data and meta learning.
ARIMA still beats the deep learning models in forecasting market data.

We develop a real time pre-processing pipeline for helping deep neural networks in pedestrian detection in low-light conditions.

Original Templates: 1,2