Interpretable Neural-Symbolic Concept Reasoning

TL;DR

This paper introduces the Deep Concept Reasoner (DCR), an interpretable neural-symbolic model that constructs and executes logical rules on concept embeddings, enhancing interpretability and performance in concept-based reasoning tasks.

Contribution

DCR is the first interpretable concept-based model that builds rule structures from concept embeddings and executes them for transparent, semantically meaningful predictions.

Findings

Improves up to +25% over state-of-the-art interpretable models.

Discovers meaningful logic rules without concept supervision.

Facilitates counterfactual generation using learned rules.

Abstract

Deep learning methods are highly accurate, yet their opaque decision process prevents them from earning full human trust. Concept-based models aim to address this issue by learning tasks based on a set of human-understandable concepts. However, state-of-the-art concept-based models rely on high-dimensional concept embedding representations which lack a clear semantic meaning, thus questioning the interpretability of their decision process. To overcome this limitation, we propose the Deep Concept Reasoner (DCR), the first interpretable concept-based model that builds upon concept embeddings. In DCR, neural networks do not make task predictions directly, but they build syntactic rule structures using concept embeddings. DCR then executes these rules on meaningful concept truth degrees to provide a final interpretable and semantically-consistent prediction in a differentiable manner. Our experiments show that DCR: (i) improves up to +25% w.r.t. state-of-the-art interpretable concept-based models on challenging benchmarks (ii) discovers meaningful logic rules matching known ground truths even in the absence of concept supervision during training, and (iii), facilitates the generation of counterfactual examples providing the learnt rules as guidance.