Learning to Solve Abstract Reasoning Problems with Neurosymbolic Program Synthesis and Task Generation

TL;DR

TransCoder is a neural program synthesis method that generates and solves abstract reasoning problems using a typed domain-specific language, synthetic task generation, and supervised training, advancing systematic learning in abstract reasoning.

Contribution

The paper introduces TransCoder, a novel neural program synthesis framework that leverages synthetic task generation and a typed language for improved abstract reasoning.

Findings

TransCoder generates tens of thousands of synthetic problems.

It achieves systematic progress in learning abstract reasoning.

The framework enables inspection and verification of solutions.

Abstract

The ability to think abstractly and reason by analogy is a prerequisite to rapidly adapt to new conditions, tackle newly encountered problems by decomposing them, and synthesize knowledge to solve problems comprehensively. We present TransCoder, a method for solving abstract problems based on neural program synthesis, and conduct a comprehensive analysis of decisions made by the generative module of the proposed architecture. At the core of TransCoder is a typed domain-specific language, designed to facilitate feature engineering and abstract reasoning. In training, we use the programs that failed to solve tasks to generate new tasks and gather them in a synthetic dataset. As each synthetic task created in this way has a known associated program (solution), the model is trained on them in supervised mode. Solutions are represented in a transparent programmatic form, which can be inspected and verified. We demonstrate TransCoder's performance using the Abstract Reasoning Corpus dataset, for which our framework generates tens of thousands of synthetic problems with corresponding solutions and facilitates systematic progress in learning.