The AI for Dynamic Systems (AIDyS (pronounced: i-dis)) focuses on Artificial Intelligence Methods for offline and online learning in dynamic systems. The primary methodology we develop is Offline and Online Reinforcement Learning for Dynamic Systems since that is currently the most promising for such systems and learning problems. There are a number of themes of interest to the lab: Offline and Online Reinforcement Learning (RL), Statistical Learning Theory for Dynamic Systems, Safe RL, Intelligent Autonomy, Multi-Agent RL, Imitation Learning and Interpretable approximations to Deep RL. Main applications of interest to the lab are autonomous robotics and vehicles, energy systems, transportation and healthcare. USC autoDrive Lab is a sister lab that focuses on problems of AI-based design and formal verification of control for intelligent autonomous systems.
Prof. Rahul Jain, Director and PI
Dhruva Mokhavinasu, Postdoc (from October 2021)
Krishna Kalagarla, RA, PhD Student
Nathan Dahlin, RA, PhD Student
Mehdi Jafarnia, RA, PhD Student
Yogesh Awate, RA, PhD Student
Rishabh Agarwal, RA, PhD Student
Akhil Agnihotri, RA, PhD Student
Cenyi Liu, MS Student researcher
William Chang, CURVE Fellow, Undergraduate researcher
Matthew Cho, CURVE Fellow, Undergraduate researcher
Boyuan Chang, Undergraduate researcher
Formal Reinforcement Learning Methods, NSF
Online Learning for real-time control of stochastic systems, NSF
New Approach to design and analysis of reinforcement algorithms, NSF