Shray Bansal
Postdoctoral Researcher
PhD, MS in Computer Science,
Georgia Institute of Technology
B.Tech. in Computer Engineering, Delhi College of Engineering
sbansal34@gatech.edu
Google Scholar / LinkedIn / CV
I am a Postdoctoral researcher at Georgia Tech working with Charles Isbell. My research is at the intersections of game theory, machine learning, multiagent systems, and human-AI coordination.
I also have a PhD from Georgia Tech, where I was advised by Charles Isbell and Ayanna Howard. My PhD research focused on developing agents (embodied and virtual) that can work alongside humans in non-zero-sum games. My work draws on research and methods from game theory, planning, reinforcement learning and Bayesian inference; and is grounded in experiments and user-studies in scenarios involving virtual agents, simulated driving and robot arms.
I also have had the opportunity to intern with some wonderful people at Microsoft Research, Honda Research Institute, and Monash University.
Publications
2024
- VIBES - Vision Backbone Efficient Selection.
- Joris Guérin, Shray Bansal, Amirreza Shaban, Paulo Mann, and Harshvardhan Gazula. arXiv, 2024.
- Reinforcement Learning with Cognitive Bias for Human-AI Ad Hoc Teamwork.
- Shray Bansal, Jin Xu, Miguel Morales, Jonathan Streater, Ayanna Howard, and Charles Isbell. CoCoMARL Workshop at the Reinforcement Learning Conference.
2023
- Game Theoretic Methods for Human-Robot Parallel Play.
- Shray Bansal. Thesis, Georgia Institute of Technology
2022
- BayesNash: Bayesian inference for Nash equilibrium selection in human-robot parallel play. [PDF]
- Shray Bansal, Jin Xu, Ayanna Howard, and Charles Isbell. Journal, Autonomous Robots.
- Nash Equilibria in Bayesian Games for Coordinating with Imperfect Humans.
- Shray Bansal, Miguel Morales, Jin Xu, Ayanna Howard, and Charles Isbell. Workshop on Strategic multi-agent interactions: game theory for robot learning and decision making at Conference on Robot Learning.
2021
- Bayesian Inference for Human-Robot Coordination in Parallel Play.
- Shray Bansal, Jin Xu, Ayanna Howard, and Charles Isbell. Workshop on Cooperative AI at NeurIPS, 2021.
2020
- Supportive Actions for Manipulation in Human-Robot Coworker Teams.
- Shray Bansal, Rhys Newbury, Wesley Chan, Akansel Cosgun, Aimee Allen, Dana Kuli´c, Tom Drummond, and Charles Isbell. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).
- Planning for Human-Robot Parallel Play via Bayesian Nash Equilibrium Inference.
- Shray Bansal, Jin Xu, Ayanna Howard, and Charles Isbell. Robotics: Science and Systems (RSS).
- Attention Driven Dynamic Memory Maps.
- Himanshu Sahni, Shray Bansal, and Charles Isbell. Workshop on Bridging AI and Cognitive Science in ICLR.
2019
- Learning to Find Common Objects Across Few Image Collections.
- Amirreza Shaban, Amir Rahimi, Shray Bansal, Stephen Gould, Byron Boots, and Richard Hartley. IEEE International Conference on Computer Vision (ICCV).
- Interaction-Aware Planning via Nash Equilibria for Manipulation in a Shared Workspace.
- Shray Bansal, Mustafa Mukadam, and Charles Isbell. Workshop on Human Movement Science for Physical Human-Robot Collaboration at ICRA.
2018
- Collaborative Planning for Autonomous Lane Merging.
- Shray Bansal, Akansel Cosgun, Alireza Nakhaei, and Kikuo Fujimura. IEEE International Conference on Intelligent Robots and Systems (IROS).
2017
- Cooperative Planning for Autonomous Lane Merging.
- Shray Bansal, Akansel Cosgun, Alireza Nakhaei, and Kikuo Fujimura. Workshop on Shared Autonomy in IROS.
- One Shot Learning for Semantic Image Segmentation.
- Amirreza Shaban, Shray Bansal, Zhen Liu, Irfan Essa, and Byron Boots. British Machine Vision Conference (BMVC).
2014
- Anticipating human actions for collaboration in the presence of task and sensor uncertainty.
- K. Hawkins, S. Bansal, N. Vo, and A. Bobick International Conference on Robotics and Automation (ICRA).
2013
- Modeling structured activity to support human-robot collaboration in the presence of task and sensor uncertainty.
- Kelsey P. Hawkins, Shray Bansal, Nam Vo, and Aaron F. Bobick. Workshop on Cognitive Robotics Systems in IROS..
- Probabilistic human action prediction and wait-sensitive planning for responsive human-robot collaboration.
- Kelsey P. Hawkins, Nam Vo, Shray Bansal, and Aaron F. Bobick. Humanoid Robots..
Teaching
I spent several semesters as a teaching assistant and head teaching assistant for the following courses at Georgia Tech:
CS 7641 Machine Learning. Professor Charles Isbell. Fall 2017 - Spring 2019.
CS 7642 Reinforcement Learning. Professor Charles Isbell. Summer 2018.
CS 6476 Computer Vision. Professor James Hays. Fall 2016.
CS 7616 Pattern Recognition. Professor Aaron Bobick. Fall 2014.
CS 4495/6495 Computer Vision. Professor Aaron Bobick. Fall 2013.
CS 3510 Design & Analysis of Algorithms. Professor Eric Vigoda. Spring 2013.