Learning STRIPS Operators from Noisy and Incomplete Observations
This addresses the issue of autonomous agents acquiring accurate domain dynamics in noisy, partial-access environments, though it appears incremental as it builds on existing STRIPS learning approaches.
The paper tackles the problem of learning STRIPS action models from noisy and incomplete observations, which is challenging in real-world domains, and shows that their method learns useful domain descriptions in simulated planning benchmarks.
Agents learning to act autonomously in real-world domains must acquire a model of the dynamics of the domain in which they operate. Learning domain dynamics can be challenging, especially where an agent only has partial access to the world state, and/or noisy external sensors. Even in standard STRIPS domains, existing approaches cannot learn from noisy, incomplete observations typical of real-world domains. We propose a method which learns STRIPS action models in such domains, by decomposing the problem into first learning a transition function between states in the form of a set of classifiers, and then deriving explicit STRIPS rules from the classifiers' parameters. We evaluate our approach on simulated standard planning domains from the International Planning Competition, and show that it learns useful domain descriptions from noisy, incomplete observations.