Koustuv Sinha
RESEARCH SCIENTIST | NEW YORK, UNITED STATES
Koustuv is a Research Scientist at Fundamental AI Research (FAIR) organization. He received his PhD from McGill University (School of Computer Science) and Mila (Quebec AI Institute), Montreal, Canada, where he was advised by Dr. Joelle Pineau. His current research interest involves investigating the role of multimodal language models at understanding and reasoning the world through rich visual representations, especially by leveraging language as a tool to decode and reason the underlying physical rules governing the world.
Research Areas
Koustuv's Publications
October 19, 2025
RESEARCH
NLP
Controlling Multimodal LLMs via Reward-guided Decoding
Aishwarya Agrawal, Adriana Romero Soriano, Koustuv Sinha, Michal Drozdzal, Oscar Mañas, Pierluca D'Oro
October 19, 2025
June 11, 2025
ROBOTICS
COMPUTER VISION
CausalVQA: A Physically Grounded Causal Reasoning Benchmark for Video Models
Aaron Foss, Ammar Rizvi, Chloe Evans, Justine T. Kao, Koustuv Sinha, Sasha Mitts
June 11, 2025
June 11, 2025
RESEARCH
COMPUTER VISION
A Shortcut-aware Video-QA Benchmark for Physical Understanding via Minimal Video Pairs
Benno Krojer, Mojtaba Komeili, Candace Ross, Koustuv Sinha, Mido Assran, Nicolas Ballas, Quentin Garrido
June 11, 2025
June 11, 2025
ROBOTICS
RESEARCH
V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning
Mojtaba Komeili, Sarath Chandar, Abha Gejji, Ada Martin, Adrien Bardes, Ammar Rizvi, Artem Zholus, Claire Roberts, Daniel Dugas, David Fan, Francisco Massa, Francois Robert Hogan, Franziska Meier, Kapil Krishnakumar, Koustuv Sinha, Marc Szafraniec, Matthew Muckley, Mido Assran, Michael Rabbat, Nicolas Ballas, Patrick Labatut, Piotr Bojanowski, Quentin Garrido, Russell Howes, Sergio Arnaud, Vasil Khalidov, Xiaodong Ma, Yann LeCun, Yong Li
June 11, 2025
December 01, 2022
NLP
CORE MACHINE LEARNING
The Curious Case of Absolute Position Embeddings
Koustuv Sinha, Adina Williams, Dieuwke Hupkes, Joelle Pineau, Amirhossein Kazemnejad, Siva Reddy
December 01, 2022
February 27, 2022
Improving Reproducibility in Machine Learning Research
Joelle Pineau, Koustuv Sinha, Alina Beygelzimer, Emily Fox, Florence d'Alché-Buc, Hugo Larochelle, Philippe Vincent-Lamarre, Vincent Larivière
February 27, 2022
September 23, 2021
Masked Language Modeling and the Distributional Hypothesis: Order Word Matters Pre-training for Little
Koustuv Sinha, Adina Williams, Dieuwke Hupkes, Douwe Kiela, Joelle Pineau, Robin Jia
September 23, 2021
November 30, 2020
HUMAN & MACHINE INTELLIGENCE
RESEARCH
Measuring Systematic Generalization in Neural Symbolic Reasoning with Transformers
Koustuv Sinha, Christopher Pal, Nicolas Gontier, Siva Reddy
November 30, 2020
June 30, 2020
June 30, 2020
June 25, 2020
RESEARCH
NLP
Learning an Unreferenced Metric for Online Dialogue Evaluation
Koustuv Sinha, Joelle Pineau, Jasmine Wang, Prasanna Parthasarathi, Ryan Lowe, William L Hamilton
June 25, 2020
November 05, 2019
RESEARCH
NLP
CLUTRR: A Diagnostic Benchmark for Inductive Reasoning from Text
Koustuv Sinha, Joelle Pineau, Will Hamilton, Jin Dong, Shagun Sodhani
November 05, 2019
Our approach
Latest news
Foundational models
Meta © 2026