Learning Algorithmic Solutions to Symbolic Planning Tasks with a Neural Computer Architecture

TL;DR

This paper introduces a neural computer architecture inspired by traditional computer designs, capable of learning and generalizing abstract algorithmic solutions for complex symbolic planning tasks using reinforcement learning.

Contribution

The paper presents a novel memory-augmented neural architecture that learns transferable algorithmic solutions across diverse symbolic planning problems.

Findings

Successfully learned solutions for Sokoban, sliding block puzzles, and robotic manipulation

Demonstrated strong generalization to larger and more complex task configurations

Achieved independence from data representation and task domain

Abstract

A key feature of intelligent behavior is the ability to learn abstract strategies that transfer to unfamiliar problems. Therefore, we present a novel architecture, based on memory-augmented networks, that is inspired by the von Neumann and Harvard architectures of modern computers. This architecture enables the learning of abstract algorithmic solutions via Evolution Strategies in a reinforcement learning setting. Applied to Sokoban, sliding block puzzle and robotic manipulation tasks, we show that the architecture can learn algorithmic solutions with strong generalization and abstraction: scaling to arbitrary task configurations and complexities, and being independent of both the data representation and the task domain.