Milan Ganai

I am a first-year PhD student in the Computer Science Department at Stanford University. I received my Bachelor of Science (2020-2023) and Master of Science (2023-2024) in Computer Science at UC San Diego, where I was advised by Professor Sicun Gao and collaborated with Professor Sylvia Herbert's Safe Autonomous Systems Lab. My interests are in Robotics, Control, and AI.

Contact: mganai at stanford dot edu

Google Scholar  |  LinkedIn

profile photo
Select Publications
Hamilton-Jacobi Reachability in Reinforcement Learning: A Survey
Milan Ganai, Sicun Gao, and Sylvia Herbert
OJ-CSYS 2024 (IEEE Open Journal of Control Systems)
arxiv | IEEE (Open Access)

A journal publication surveying the recent literature on scalable Hamilton-Jacobi reachability estimation in reinforcement learning to provide a foundational basis for research into reliability in high-dimensional systems. We review how this technique has been employed to solve challenging tasks like those with dynamic obstacles and lidar-based or RGB image-based observations.

Iterative Reachability Estimation for Safe Reinforcement Learning
Milan Ganai, Zheng Gong, Chenning Yu, Sylvia Herbert, and Sicun Gao
NeurIPS 2023 (Conference on Neural Information Processing Systems)
paper | openreview | code | website

Hamilton-Jacobi reachability estimation for model-free safe RL in deterministic and stochastic environments with safety guarantees and convergence analysis. Tasks include lidar-based observations, dynamic obstacles, and multiple hard and soft constraints.

Learning Stabilization Control from Observations by Learning Lyapunov-like Proxy Models
Milan Ganai, Chiaki Hirayama, Ya-Chien Chang, and Sicun Gao
ICRA 2023 (IEEE International Conference on Robotics and Automation)
paper | IEEE | website

Learning Lyapunov-like models offline from observation-only, expert data to solve stabilization control tasks online. Deployed in hardware for robustness testing.

Target-independent XLA optimization using Reinforcement Learning
Milan Ganai, Haichen Li, Theodore Enns, Yida Wang, and Randy Huang
ML for Systems @ NeurIPS 2022 (Workshop on ML for Systems in Neural Information Processing Systems)
paper | website

Reinforcement Learning to determine XLA compiler optimization pass ordering to reduce GPT-2, BERT, and ResNet graph sizes.

Identifying Merged Tracks in Dense Environments with Machine Learning
Patrick McCormack, Milan Ganai, Ben Nachman, and Maurice Garcia-Sciveres
CTD/WIT 2019 (Connecting the Dots / Workshop on Intelligent Trackers)
paper

Building boosted decision trees to classify reconstructed particle tracks as merged in high density particle physics environments.



Academic Services

Conference Reviewer: ICRA 2023, L4DC 2024, ICML 2024, AAAI 2025, ICLR 2025


Original code template