GeezerVison

GeezerVison is showing bsarkar321.github.io — Bidipta Sarkar

Bidipta Sarkar

I am a second-year DPhil student in Engineering Science at the University of Oxford in the FLAIR and WhiRL labs, co-supervised by Professor Jakob Foerster and Professor Shimon Whiteson. I am funded by the Clarendon Fund Scholarship in partnership with a Department of Engineering Science Studentship.

I received my BS in Computer Science with Honors and Distinction at Stanford University (2020-2024), where I was also a member of Professor Dorsa Sadigh's ILIAD lab since my sophomore year.

I am interested in creating AI agents that can interact with their environment and safely work alongside humans and other autonomous agents, with a growing focus on integrating natural language into AI coordination. My research broadly spans three subfields of computer science:

• Multi-Agent Reinforcement Learning: Enabling independently trained agents to cooperate on a common task and form conventions.

• Vision: Capturing meaningful information about an agent's environment from sensors.

• Graphics: Simulating environments while balancing speed and realism.

Posts

(December 2025, Livestream) EGGROLL Discussion with Yacine Mahdid

(May 2025) Easy Academic Website Template (With Instructions)

(November 2023, Video) My talk on Diverse Conventions for Human-AI Collaboration

(August 2023, Blog Post) Overcooked in Thousands of Kitchens: Training Top Performing Agents in Under a Minute

(Living Document) Emacs Setup for macOS and GNU/Linux

Publications

Learning to Drive in New Cities Without Human Demonstrations

Zilin Wang*, Saeed Rahmani*, Daphne Cornelisse, Bidipta Sarkar, Alexander David Goldie, Jakob Nicolaus Foerster*, Shimon Whiteson*

arXiv preprint, February 2026

Paper / Website / Code

Evolution Strategies at the Hyperscale

Bidipta Sarkar*, Mattie Fellows*, Juan Agustin Duque*, Alistair Letcher\(^\dagger\), Antonio León Villares\(^\dagger\), Anya Sims\(^\dagger\), Clarisse Wibault\(^\dagger\), Dmitry Samsonov\(^\dagger\), Dylan Cope\(^\dagger\), Jarek Liesen\(^\dagger\), Kang Li\(^\dagger\), Lukas Seier\(^\dagger\), Theo Wolf\(^\dagger\), Uljad Berdica\(^\dagger\), Valentin Mohl\(^\dagger\), Alexander David Goldie, Aaron Courville, Karin Sevegnani, Shimon Whiteson*, Jakob Nicolaus Foerster*

International Conference on Machine Learning (ICML), July 2026
arXiv preprint, November 2025

Paper / Website / Code / Nano-EGG Code

Discrete Flow Matching is a Surprisingly Effective Post-training Method to Address Compound Error in Autoregressive Models

Kang Li*, Bidipta Sarkar*, Zheng Xiong*, Sascha Frey, Zilin Wang, Frensi Zejnullahu, Alfred Backhouse, Stefan Zohren, Anisoara Calinescu, Mihai Cucuringu\(^\dagger\), Jakob Foerster\(^\dagger\)

ACM International Conference on AI in Finance (Oral), November 2025

Paper

LOB-Bench: Benchmarking Generative AI for Finance – an Application to Limit Order Book Data

Peer Nagy*, Sascha Frey*, Kang Li, Bidipta Sarkar, Svitlana Vyetrenko, Stefan Zohren, Ani Calinescu, Jakob Foerster

International Conference on Machine Learning (ICML), July 2025

Paper / Website / Code

Training Language Models for Social Deduction with Multi-Agent Reinforcement Learning

Bidipta Sarkar, Warren Xia, C. Karen Liu, Dorsa Sadigh

International Conference on Autonomous Agents and Multiagent Systems (AAMAS) (Oral), May 2025
Stanford Senior Honors Thesis, May 2024

Paper / Stanford Digital Repository / Website / Code / Models / Poster

Physically Grounded Vision-Language Models for Robotic Manipulation

Jensen Gao, Bidipta Sarkar, Fei Xia, Ted Xiao, Jiajun Wu, Brian Ichter, Anirudha Majumdar, Dorsa Sadigh

International Conference on Robotics and Automation (ICRA), May 2024

Paper / Website / Video

Diverse Conventions for Human-AI Collaboration

Bidipta Sarkar, Andy Shih, Dorsa Sadigh

Conference on Neural Information Processing Systems (NeurIPS), December 2023

Paper / Website / Code / Video / Poster

An Extensible, Data-Oriented Architecture for High-Performance, Many-World Simulation

Brennan Shacklett, Luc Guy Rosenzweig, Zhiqiang Xie, Bidipta Sarkar, Andrew Szot, Erik Wijmans, Vladlen Koltun, Dhruv Batra, Kayvon Fatahalian

Transactions on Graphics 2023

Paper / Website / RL Environments / Blog / Colab

PantheonRL: A MARL Library for Dynamic Training Interactions

Bidipta Sarkar*, Aditi Talati*, Andy Shih*, Dorsa Sadigh

Proceedings of the 36th AAAI Conference on Artificial Intelligence (Demo Track), February 2022

Paper / Code / Video / DOI / Poster

Workshop Papers

Zero-Shot Coordination among LLM Agents

Adrian Hayler*, Shashank Reddy Chirra*, Andrei Lupu, Johannes Forkel, Bidipta Sarkar, Siheng Feng, Jakob Nicolaus Foerster

ICLR Workshop on Multi-Agent Learning and Its Opportunities in the Era of Generative AI, April 2026

Paper

An Interactive Agent Foundation Model

Zane Durante*, Bidipta Sarkar*, Ran Gong*, Rohan Taori, Yusuke Noda, Paul Tang, Ehsan Adeli, Shrinidhi Kowshika Lakshmikanth, Kevin Schulman, Arnold Milstein, Demetri Terzopoulos, Ade Famoti, Noboru Kuno, Ashley Llorens, Hoi Vo, Katsu Ikeuchi, Li Fei-Fei, Jianfeng Gao, Naoki Wake*, Qiuyuan Huang*

Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR) Workshops, June 2025

Paper / Website