|
I am a first year Ph.D. student at Mila and Université de Montréal where I am advised by Prof. Aaron Courville. Previously, I finished my Masters at Mila and UdeM under the supervision of Prof. Yoshua Bengio and Dr. Anirudh Goyal. My recent contributions have revolved around improving and evaluating the reasoning capabilities of large language models (LLMs) through RL based post-training of LLMs, synthetic data generation and test-time scaling. I am broadly interested in investigating self-improvement in LLM reasoning and how these advancements can be leveraged for open ended mathematical discovery. My previous work involved incorporating cognitively inspired inductive biases into deep learning models for low-compute unlearning and for improving out of distribution generalization in multi-agent reinforcement learning. I completed my undergraduate studies in Electronics and Communications Engineering from BITS Pilani, Goa Campus where I worked with 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. |
|
|
|
|
|
For a complete list of publications, please visit my Google Scholar page. |
|
Siddarth Venkataraman*, Vineet Jain*, Sarthak Mittal*, Vedant Shah, Johan Obando-Ceron, Yoshua Bengio, Brian Bartoldson, Bhavya Kailkhura, Guillaume Lajoie, Glen Berseth, Nikolay Malkin, Moksh Jain Pre-print arXiv / Website Using recursive aggregation of the chain of thought (CoT) of LLMs to unlock powerful test-time scaling capabilities. |
|
Vedant Shah, Dingli Yu, Kaifeng Lyu, Simon Park, Jiatong Yu, Yinghui He, Nan Rosemary Ke, Michael Mozer, Yoshua Bengio, Sanjeev Arora, Anirudh Goyal MATH-AI Workshop, NeurIPS 2024 arXiv / Website / Poster We use frontier LLMs such as GPT-4 and Claude 3 and Gemini to generate challenging mathematical questions by asking them to compose two domain skills at once. Based on this, we present a new math eval where a large number of open source as well as proprietary LLMs show significant drops relative to a standard eval. |
|
Vedant Shah, Frederik Träuble, Ashish Malik, Hugo Larochelle, Michael Mozer, Sanjeev Arora, Yoshua Bengio, Anirudh Goyal TMLR 2025 Openreview We propose a nearly compute-free approach for class unlearning in multi-class classification settings. |
|
Dianbo Liu*, Vedant Shah*, Oussama Boussif*, Cristian Meo, Anirudh Goyal, Tianmin Shu, Michael Mozer, Nicolas Heess, Yoshua Bengio, ICLR 2023 arXiv / Talk Investigating the problem of environmental heterogeneity and using behvioural priors to tackle the problems of coordination and heterogeneity in MARL |
|
Thomas Jiralerspong, Xiaoyin Chen, Yash More, Vedant Shah, Yoshua Bengio How Far are We From AGI? Workshop, ICLR 2024 arXiv / Poster We propose an approach to discover full causal graphs in real-world settings using LLMs such at GPT-4 by augmenting LLM querying with a BFS like procedure. |
|
Michał Koziarski, Mohammed Abukalam, Vedant Shah, Louis Vaillancourt, Doris Alexandra Schuetz, Moksh Jain, Almer van der Sloot, Mathieu Bourgey, Anne Marinier, Yoshua Bengio GEM Workshop, ICLR 2024 arXiv / Poster We use PPI modulation task based reward models to train GFlowNets to address the combinatorially challenging task of generating DNA Encoded Libraries (DELs) |
|
|
 |
Sep 2023 - Present
Graduate Student Researcher Advisor - Dr. Anirudh Goyal and Prof. Yoshua Bengio Sep 2021 - July 2023 Research Intern Advisors - Dr. Anirudh Goyal and Prof. Yoshua Bengio |
 |
Jan 2022 - Aug 2022
Undergraduate Researcher Advisor - Prof. Ashwin Srinivasan Jan 2021 - Aug 2021 Undergraduate Researcher Advisor - Prof. Tanmay Verlekar |
 |
May 2021 - Sep 2021
Research Intern Advisor - Dr. Gautam Shroff |
 |
May 2021 - Aug 2021
Student Developer Organisation - GFOSS (Suborganisation - OpenDR) |
 |
Aug. 2020 - Nov. 2020
Research Intern Advisor - Prof. GC Nandi |
|
|