Towards Learning Abductive Reasoning using VSA Distributed Representations

TL;DR

This paper introduces ARLC, a novel abductive reasoning model using VSA distributed representations, achieving state-of-the-art accuracy on Raven's matrices with improved interpretability and robustness, surpassing larger models.

Contribution

ARLC is a new model that combines programming domain knowledge with learned rules, offering a broadly applicable training objective for abductive reasoning.

Findings

ARLC achieves state-of-the-art accuracy on I-RAVEN dataset.

ARLC outperforms larger neuro-symbolic and connectionist models.

ARLC demonstrates robustness to incremental learning without catastrophic forgetting.

Abstract

We introduce the Abductive Rule Learner with Context-awareness (ARLC), a model that solves abstract reasoning tasks based on Learn-VRF. ARLC features a novel and more broadly applicable training objective for abductive reasoning, resulting in better interpretability and higher accuracy when solving Raven's progressive matrices (RPM). ARLC allows both programming domain knowledge and learning the rules underlying a data distribution. We evaluate ARLC on the I-RAVEN dataset, showcasing state-of-the-art accuracy across both in-distribution and out-of-distribution (unseen attribute-rule pairs) tests. ARLC surpasses neuro-symbolic and connectionist baselines, including large language models, despite having orders of magnitude fewer parameters. We show ARLC's robustness to post-programming training by incrementally learning from examples on top of programmed knowledge, which only improves its performance and does not result in catastrophic forgetting of the programmed solution. We validate ARLC's seamless transfer learning from a 2x2 RPM constellation to unseen constellations. Our code is available at https://github.com/IBM/abductive-rule-learner-with-context-awareness.