Haque Ishfaq

I am a final year PhD Student at Mila and McGill University, advised by Prof. Doina Precup. My current research focuses on exploration in reinforcement learning, large language model alignment and reasoning.

I am on the job market and actively looking for Research Scientist or Postdoctoral Researcher positions in the USA/Canada!

Previously, I obtained my BS (Mathematical and Computational Science) in 2016 and MS (Statistics) in 2018 from Stanford University. Previously, I also did research internships at Meta AI, Microsoft, IBM Research and Nvidia.

Feel free to reach out to me in case you have any questions or want to chat about my work!

I co-organize the RLHF reading group at Mila. If you want to present your work, feel free to reach out!

In my high school days, I was heavily involved with mathematical olympiads and represented Bangladesh 3 times at the International Mathematical Olympiad (IMO).

Email / CV / Google Scholar / Github / Twitter / LinkedIn

News

January 2025: One paper accepted to ICLR 2025.
September 2024: One paper accepted to NeurIPS 2024.
May 2024: One paper accepted to Reinforcement Learning Conference (RLC) 2024.
May 2024: Invited to attend and present at the 2024 Two Sigma PhD Symposium in NYC, USA.
April 2024: Invited to give a long talk at the 25th International Symposium on Mathematical Programming, Montreal, Canada.
January 2024: One paper accepted to ICLR 2024.
April 2023: Invited to attend and present at the Citadel PhD Summit in Miami, USA.
Summer 2022: Research internship at Meta AI with Applied Reinforcement Learning Team
September 2021: Invited talk at ML-Seminar, Purdue University. Host: Kamyar Azizzadenesheli
September 2021: Invited talk at Jiantao Jiao Group meeting, University of California, Berkeley. Host: Jiantao Jiao
February 2021: Invited talk at Benjamin Van Roy Group meeting, Stanford University. Host: Benjamin Van Roy
January 2021: Contributed talk at Microsoft Research Reinforcement Learning Day 2021.
November 2020: Contributed talk at Northeast Reinforcement Learning and Decision Making Symposium.

Conference Publications

Langevin Soft Actor-Critic: Efficient Exploration through Uncertainty-Driven Critic Learning
Haque Ishfaq*, Guangyuan Wang*, Sami Nur Islam and Doina Precup
International Conference on Learning Representations (ICLR), 2025
[Paper], [Code]

Offline Multitask Representation Learning for Reinforcement Learning
Haque Ishfaq, Thanh Nguyen-Tang, Songtao Feng, Raman Arora, Mengdi Wang, Ming Yin and Doina Precup
Conference on Neural Information Processing Systems (NeurIPS), 2024
[Paper], [Poster]

More Efficient Randomized Exploration for Reinforcement Learning via Approximate Sampling
Haque Ishfaq*, Yixin Tan*, Yu Yang, Qingfeng Lan, Jianfeng Lu, A. Rupam Mahmood, Doina Precup and Pan Xu
Reinforcement Learning Conference (RLC), 2024
[Paper], [Code], [Poster]

Provable and Practical: Efficient Exploration in Reinforcement Learning via Langevin Monte Carlo
Haque Ishfaq*, Qingfeng Lan*, Pan Xu, A. Rupam Mahmood, Doina Precup, Anima Anandkumar and Kamyar Azizzadenesheli
International Conference on Learning Representations (ICLR), 2024
[Paper], [Code], [Poster]

Randomized Exploration for Reinforcement Learning with General Value Function Approximation
Haque Ishfaq*, Qiwen Cui*, Viet Nguyen, Alex Ayoub, Zhuoran Yang, Zhaoran Wang, Doina Precup, and Lin F. Yang
International Conference on Machine Learning (ICML), 2021
[Paper], [Slide]

Preprints/Workshop Publications

Randomized Least Squares Policy Optimization
Haque Ishfaq, Zhuoran Yang, Andrei Lupu, Viet Nguyen, Lewis Liu, Riashat Islam, Zhaoran Wang and Doina Precup
ICML Workshop on Reinforcement Learning Theory , 2021
[Paper]

Path-Based Contextualization of Knowledge Graphs for Textual Entailment
Kshitij Fadnis, Kartik Talamadupula, Pavan Kapanipathi, Haque Ishfaq, Salim Roukos and Achille Fokoue
Preprint, 2019
[Paper]

TVAE: Triplet-Based Variational Autoencoder using Metric Learning
Haque Ishfaq, Assaf Hoogi and Daniel Rubin
Preprint, 2018
[Paper]

Website template