AMLSI: A Novel Accurate Action Model Learning Algorithm
This work addresses the problem of learning accurate PDDL action models for automated planning, which is crucial for enabling planners to operate effectively in unknown environments.
This paper introduces AMLSI, a grammar induction-based algorithm that learns PDDL action models through trial and error by observing system state transitions from randomly generated action sequences. AMLSI demonstrates the ability to learn domains with sufficient accuracy from partial and noisy observations, enabling planners to solve new problems.
This paper presents new approach based on grammar induction called AMLSI Action Model Learning with State machine Interactions. The AMLSI approach does not require a training dataset of plan traces to work. AMLSI proceeds by trial and error: it queries the system to learn with randomly generated action sequences, and it observes the state transitions of the system, then AMLSI returns a PDDL domain corresponding to the system. A key issue for domain learning is the ability to plan with the learned domains. It often happens that a small learning error leads to a domain that is unusable for planning. Unlike other algorithms, we show that AMLSI is able to lift this lock by learning domains from partial and noisy observations with sufficient accuracy to allow planners to solve new problems.