Improving Safety in Reinforcement Learning Using Model-Based Architectures and Human Intervention

This work is supported by U.S. Army Research Laboratory under Cooperative Agreement Number W911NF-10-2-0022