Safe Learning of PDDL Domains with Conditional Effects

Argaman Mordoch, Enrico Scala, Roni Stern, Brendan Juba

Keywords: Automated Planning, Action Model Learning, Conditional Effects, Safety, Inductive Learning

TL;DR: A novel approach to learning action models with conditional and universal effects while maintaining safety.

Abstract:

Powerful domain-independent planners have been developed to solve various types of planning problems.
These planners often require a model of the acting agent's actions, given in some planning domain description language.
Manually designing such an action model is a notoriously challenging task.
An alternative is to automatically learn action models from observation.
Such an action model is called safe if plans consistent with it are also consistent with the real, unknown action model.
Algorithms for learning such safe action models exist, yet they cannot handle domains with conditional or universal effects, which are common constructs in many planning problems.
We prove that learning non-trivial safe action models with conditional effects may require an exponential number of samples.
Then, we identify reasonable assumptions under which such learning is tractable and propose Conditional-SAM, the first algorithm capable of doing so.
We analyze this Conditional-SAM theoretically and evaluate it experimentally.
Our results show that the action models learned by Conditional-SAM can be used to solve perfectly most of the test set problems in most of the experimented domains.

Primary Keywords: Learning, Knowledge Representation/Engineering

Category: Long

Student: Graduate

Submission Number: 239