Trained Policy

Trained policy research focuses on developing robust and efficient methods for transferring policies learned in simulation to real-world robotic and control systems, bridging the "reality gap." Current efforts concentrate on improving policy gradient methods, incorporating model-based reinforcement learning for enhanced autonomy and data efficiency, and leveraging techniques like prompt learning and Bayesian controller fusion to enhance sim-to-real transfer and handle uncertainty. This research is crucial for accelerating the adoption of reinforcement learning in real-world applications, particularly in robotics and complex dynamic systems, by reducing the reliance on extensive real-world data collection and improving the reliability of deployed agents.

Papers

September 1, 2024

Diffusion Policy Policy Optimization
Allen Z. Ren, Justin Lidard, Lars L. Ankile, Anthony Simeonov, Pulkit Agrawal, Anirudha Majumdar, Benjamin Burchfiel, Hongkai Dai, Max Simchowitz
Policy Gradient Diffusion Policy Trained Policy

August 19, 2024

World Models Increase Autonomy in Reinforcement Learning
Zhao Yang, Thomas M. Moerland, Mike Preuss, Aske Plaat, Edward S. Hu
Reinforcement Learning World Model Policy Learning Trained Policy Free Task

October 19, 2023

SDGym: Low-Code Reinforcement Learning Environments using System Dynamics Models
Emmanuel Klu, Sameer Sethi, DJ Passey, Donald Martin Jr
Reinforcement Learning Reinforcement Learning Environment System Dynamic Trained Policy

August 28, 2023

Prompt to Transfer: Sim-to-Real Transfer for Traffic Signal Control with Prompt Learning
Longchao Da, Minquan Gao, Hao Mei, Hua Wei
Reinforcement Learning Prompt Learning Formality Transfer Style PROMPT Traffic Signal Control DQN Agent Traffic Dynamic Trained Policy

March 4, 2022

Cloud-Edge Training Architecture for Sim-to-Real Deep Reinforcement Learning
Hongpeng Cao, Mirco Theile, Federico G. Wyrwal, Marco Caccamo
Deep Reinforcement Learning Sim to Real Cloud Based Sim to Real Reinforcement Learning Trained Policy

February 11, 2022

Uncertainty Aware System Identification with Universal Policies
Buddhika Laknath Semage, Thommen George Karimpanal, Santu Rana, Svetha Venkatesh
Reinforcement Learning Robust Policy Domain Randomization Uncertain System Robust Bayesian Optimization Trained Policy

December 10, 2021

Zero-Shot Uncertainty-Aware Deployment of Simulation Trained Policies on Real-World Robots
Krishan Rana, Vibhavari Dasagi, Jesse Haviland, Ben Talbot, MIchael Milford, Niko Sünderhauf
Deep Reinforcement Learning Real Robot Continuous Control Task Uncertainty Aware Deployment Laboratory Robotics Trained Policy