MADIL: An MDL-based Framework for Efficient Program Synthesis in the ARC Benchmark

TL;DR

MADIL is an MDL-based framework for program synthesis in the ARC benchmark, emphasizing efficiency and interpretability over raw performance, and demonstrating structured generalization through pattern decomposition.

Contribution

Introduces MADIL, a novel MDL-based approach for efficient and interpretable program synthesis in the ARC benchmark, contrasting with resource-intensive LLM methods.

Findings

MADIL achieves 7% at ArcPrize 2024.

MADIL offers greater efficiency than LLM-based methods.

MADIL enables structured generalization through pattern decomposition.

Abstract

Artificial Intelligence (AI) has achieved remarkable success in specialized tasks but struggles with efficient skill acquisition and generalization. The Abstraction and Reasoning Corpus (ARC) benchmark evaluates intelligence based on minimal training requirements. While Large Language Models (LLMs) have recently improved ARC performance, they rely on extensive pre-training and high computational costs. We introduce MADIL (MDL-based AI), a novel approach leveraging the Minimum Description Length (MDL) principle for efficient inductive learning. MADIL performs pattern-based decomposition, enabling structured generalization. While its performance (7% at ArcPrize 2024) remains below LLM-based methods, it offers greater efficiency and interpretability. This paper details MADIL's methodology, its application to ARC, and experimental evaluations.