Reward is not Necessary: How to Create a Compositional Self-Preserving Agent for Life-Long Learning.
Thomas J. RingstromPublished in: CoRR (2022)
Keyphrases
- life long learning
- multi agent
- multiagent systems
- reward function
- lifelong learning
- learning agent
- multi agent systems
- intelligent agents
- autonomous agents
- e learning
- dynamic environments
- software agents
- life long
- reinforcement learning
- mobile agents
- decision making
- initially unknown
- agent receives
- multiple agents
- long run
- action selection
- inverse reinforcement learning
- agent technology
- expected reward
- learning materials
- state space