1 paper · 1 filter
Michel Ma, Takuma Seno, Kaushik Subramanian +3
When designing reinforcement learning (RL) agents, a designer communicates the desired agent behavior through the definition of reward functions - numerical feedback given to the a…