February 17, 2021 Artificial intelligence Leave a comment

Designer-centered reinforcement learning

An animation titled “Adding style to RL agent navigation” plays through three versions of agent behavior in a navigation task. Each shows the agent, represented by a robot icon, moving around two walls to reach a goal, represented by a trophy icon. The first version applies only a task reward, and the agent path to the goal is too close to the wall. The second applies an excess shaping reward, and the path is too far from the wall. The third balances reward components, and the path is more central.

In video games, nonplayer characters, bots, and other game agents help bring a digital world and its story to life. They can help make the mission of saving humanity feel urgent,

To finish reading, please visit source site

Leave a Reply Cancel reply

You must be logged in to post a comment.