How RL Agents Behave When Their Actions Are Modified
By Eric Langlois, Tom Everitt. 2021
19 citations recorded by the source corpus.
Reinforcement learning in complex environments may require supervision to prevent the agent from attempting dangerous actions. As a result of supervisor intervention, the executed action may differ from the action specified by the policy. How does this affect learning? We present the Modified-Action Markov Decision Process, an extension of the MDP model that allows actions to differ from the policy. We analyze the asymptotic behaviours of common reinforcement learning algorithms in this setting and show that they adapt in different ways: some completely ignore modifications while others go to various lengths in trying to avoid action modifications that decrease reward. By choosing the right algorithm, developers can prevent their agents from learning to circumvent interruptions or constraints, and better control agent responses to other kinds of action modification, like self-damage.
Subfields: Artificial Intelligence, Information Systems
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Matthew Riemer
- Tom Everitt
- Irina Rish
- Khimya Khetarpal
- Shane Legg
- Doina Precup
- Pedro A. Ortega
- Eric Langlois
In Information Systems
Citations in the corpus
- Towards Continual Reinforcement Learning: A Review and Perspectives (2020, background): “While less common, work on non-stationary RL has also considered the case of a changing action space (Chandak, Theocharous, Kostas, Jordan, and Thomas, 2019; Langlois and Everitt, 2021; Jain, Kosaka, Kim, and Lim, 2021; Trabucco, Phielipp, and Berseth, 2022) and observation function (Trabucco et al.”
- Agent Incentives: A Causal Perspective (2021, background): “…diagrams have already aided in the understanding of incentive problems such as an agent’s redirectability, ambition, tendency to tamper with reward, and other properties (Armstrong et al. 2020; Holtman 2020; Cohen, Vellambi, and Hutter 2020; Everitt et al. 2019a,b; Langlois and Everitt 2021).”