Sumeet Batra

I am a 5th year PhD candidate at the University of Southern California advised by Prof. Gaurav Sukhatme in the Robotics Embedded Systems Lab (RESL). My research lies at the intersection of curiosity-driven exploration, perception, and decision-making for robots – focused on how embodied agents can explore, understand, and act in the world in a generalizable manner. I’m particularly interested in how data-driven models can achieve extrapolative generalization: enabling robots to generalize beyond their training data by learning world models and representations that support robust decision-making in novel environments. To this end, I’ve published work on Reinforcement Learning (RL), open-ended learning algorithms such as Quality Diversity (QD) for self-supervised exploration, Generative models such as Diffusion models for robotics, and self-supervised representation learning from RGB inputs for embodied agents.

Previously, I’ve had the privilege of working with the Autonomous Vehicles team at NVIDIA as a research scientist intern in 2022 and 2023 on Reinforcement Learning for autonomous driving and Diffusion models for scenario generation. Before that, I was an intern at the National Institute of Standards and Technology (NIST), where I worked on Generative Adversarial Networks (GANs).