Learning Type-Generalized Actions for Symbolic Planning

TL;DR

This paper introduces a method to learn and generalize symbolic actions in planning tasks using entity hierarchies, enabling transferability and adaptation to new, complex scenarios with minimal observations.

Contribution

It proposes a novel approach to learn type-generalized actions from observations, improving transferability and adaptability in symbolic planning.

Findings

Type-generalized actions can be learned from few observations.

The approach generalizes to novel situations with longer sequences and new entities.

On-the-fly generalization improves planning in unexpected environments.

Abstract

Symbolic planning is a powerful technique to solve complex tasks that require long sequences of actions and can equip an intelligent agent with complex behavior. The downside of this approach is the necessity for suitable symbolic representations describing the state of the environment as well as the actions that can change it. Traditionally such representations are carefully hand-designed by experts for distinct problem domains, which limits their transferability to different problems and environment complexities. In this paper, we propose a novel concept to generalize symbolic actions using a given entity hierarchy and observed similar behavior. In a simulated grid-based kitchen environment, we show that type-generalized actions can be learned from few observations and generalize to novel situations. Incorporating an additional on-the-fly generalization mechanism during planning, unseen task combinations, involving longer sequences, novel entities and unexpected environment behavior, can be solved.