An Accurate HDDL Domain Learning Algorithm from Partial and Noisy Observations
This addresses the problem of error-prone manual domain specification for HTN planning, which is more complex than classical planning, but the approach is incremental as it builds on existing learning methods.
The authors tackled the difficulty of manually encoding Hierarchical Task Network (HTN) planning domains by proposing HierAMLSI, a grammar induction-based approach that learns action models and HTN methods from partial and noisy observations, achieving high accuracy.
The Hierarchical Task Network ({\sf HTN}) formalism is very expressive and used to express a wide variety of planning problems. In contrast to the classical {\sf STRIPS} formalism in which only the action model needs to be specified, the {\sf HTN} formalism requires to specify, in addition, the tasks of the problem and their decomposition into subtasks, called {\sf HTN} methods. For this reason, hand-encoding {\sf HTN} problems is considered more difficult and more error-prone by experts than classical planning problem. To tackle this problem, we propose a new approach (HierAMLSI) based on grammar induction to acquire {\sf HTN} planning domain knowledge, by learning action models and {\sf HTN} methods with their preconditions. Unlike other approaches, HierAMLSI is able to learn both actions and methods with noisy and partial inputs observation with a high level or accuracy.