Scaling up ML-based Black-box Planning with Partial STRIPS Models

TL;DR

This paper proposes using incomplete STRIPS models to enhance ML-based black-box planning by leveraging relaxation heuristics, improving planning efficiency without requiring full symbolic models.

Contribution

Introducing a method to incorporate partial STRIPS models into ML-based planning to guide search heuristics when full models are unavailable.

Findings

Partial STRIPS models improve planning efficiency

Relaxation heuristics guide search effectively

Method outperforms data collection and tuning approaches

Abstract

A popular approach for sequential decision-making is to perform simulator-based search guided with Machine Learning (ML) methods like policy learning. On the other hand, model-relaxation heuristics can guide the search effectively if a full declarative model is available. In this work, we consider how a practitioner can improve ML-based black-box planning on settings where a complete symbolic model is not available. We show that specifying an incomplete STRIPS model that describes only part of the problem enables the use of relaxation heuristics. Our findings on several planning domains suggest that this is an effective way to improve ML-based black-box planning beyond collecting more data or tuning ML architectures.