Article: Knowledge graph enhanced interpretable interlanguage error correction model Journal: International Journal of Reasoning-based Intelligent Systems (IJRIS) 2026 Vol.18 No.21 pp.47 - 59 Abstract: Grammatical error correction plays a crucial role in second language acquisition, especially in assisting English interlanguage learners. In this paper, we propose a novel knowledge graph enhanced interpretable English interlanguage error correction model. Specifically, we construct an interlanguage knowledge graph that maps lexical, syntactic, and semantic error patterns onto standard English grammar rules. Context-aware sequence-to-sequence architectures combined with graph fusion algorithms dynamically incorporate this structured external knowledge. Furthermore, the interpretability generation module retrieves the target rationale to form explicit linguistic feedback. Extensive experiments on the BEA-2019 benchmark and the constructed learner corpus show that our model outperforms existing baseline models by a true reasonable margin of 1.8%. At the same time, the retrieval accuracy reaches 85.6%, which confirms that the framework significantly enhances structural understanding and teaching transparency in automatic error correction. Inderscience Publishers - linking academia, business and industry through research

Title: Knowledge graph enhanced interpretable interlanguage error correction model

Authors: Xiaoyan Cao

Addresses: School of Liberal Arts, Wuxi Taihu University, Wuxi, 214063, China

Abstract: Grammatical error correction plays a crucial role in second language acquisition, especially in assisting English interlanguage learners. In this paper, we propose a novel knowledge graph enhanced interpretable English interlanguage error correction model. Specifically, we construct an interlanguage knowledge graph that maps lexical, syntactic, and semantic error patterns onto standard English grammar rules. Context-aware sequence-to-sequence architectures combined with graph fusion algorithms dynamically incorporate this structured external knowledge. Furthermore, the interpretability generation module retrieves the target rationale to form explicit linguistic feedback. Extensive experiments on the BEA-2019 benchmark and the constructed learner corpus show that our model outperforms existing baseline models by a true reasonable margin of 1.8%. At the same time, the retrieval accuracy reaches 85.6%, which confirms that the framework significantly enhances structural understanding and teaching transparency in automatic error correction.

Keywords: knowledge graph; grammatical error correction; interlanguage; explainable AI; natural language processing; NLP.

DOI: 10.1504/IJRIS.2026.155784

International Journal of Reasoning-based Intelligent Systems, 2026 Vol.18 No.21, pp.47 - 59

Received: 05 Jun 2026
Accepted: 05 Jul 2026
Published online: 13 Aug 2026 *

Title: Knowledge graph enhanced interpretable interlanguage error correction model

Keep up-to-date