CLMN: Concept based Language Models via Neural Symbolic Reasoning

View PDF HTML (experimental)

Abstract:Deep learning has advanced NLP, but interpretability remains limited, especially in healthcare and finance. Concept bottleneck models tie predictions to human concepts in vision, but NLP versions either use binary activations that harm text representations or latent concepts that weaken semantics, and they rarely model dynamic concept interactions such as negation and context. We introduce the Concept Language Model Network (CLMN), a neural-symbolic framework that keeps both performance and interpretability. CLMN represents concepts as continuous, human-readable embeddings and applies fuzzy-logic reasoning to learn adaptive interaction rules that state how concepts affect each other and the final decision. The model augments original text features with concept-aware representations and automatically induces interpretable logic rules. Across multiple datasets and pre-trained language models, CLMN achieves higher accuracy than existing concept-based methods while improving explanation quality. These results show that integrating neural representations with symbolic reasoning in a unified concept space can yield practical, transparent NLP systems.

Comments: 7 pages, 2 figures

Subjects:

Computation and Language (cs.CL); Artificial Intelligence (cs.AI)

MSC classes: 68T50, 68T07, 68T27

ACM classes: I.2.7; I.2.6

Cite as: arXiv:2510.10063 [cs.CL]

(or arXiv:2510.10063v2 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2510.10063

arXiv-issued DOI via DataCite

Submission history

From: Yang Yibo [view email] [v1] Sat, 11 Oct 2025 06:58:44 UTC (523 KB) [v2] Mon, 30 Mar 2026 16:44:05 UTC (739 KB)