TL;DR
This paper introduces a neuro-symbolic reasoning architecture that combines neural perception with symbolic transformation filtering, significantly improving generalization on the ARC task without task-specific training.
Contribution
It proposes a novel compositional neuro-symbolic framework that enhances language models with object-based reasoning, achieving better generalization on structured reasoning tasks.
Findings
Performance on ARC-AGI-2 improved from 16% to 24.4%.
Combining with ARC Lang Solver increases accuracy to 30.8%.
The approach reduces reliance on brute-force search and sampling.
Abstract
We study structured abstraction-based reasoning for the Abstraction and Reasoning Corpus (ARC) and compare its generalization to test-time approaches. Purely neural architectures lack reliable combinatorial generalization, while strictly symbolic systems struggle with perceptual grounding. We therefore propose a neuro-symbolic architecture that extracts object-level structure from grids, uses neural priors to propose candidate transformations from a fixed domain-specific language (DSL) of atomic patterns, and filters hypotheses using cross-example consistency. Instantiated as a compositional reasoning framework based on unit patterns inspired by human visual abstraction, the system augments large language models (LLMs) with object representations and transformation proposals. On ARC-AGI-2, it improves base LLM performance from 16% to 24.4% on the public evaluation set, and to 30.8% when…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
