Read the full-text of this article for free

Title: Optimising college English translation course teaching strategies by multi-modal representation learning

Authors: Ru Wang

Addresses: School of General Education, Hainan Vocational University, Haikou, Hainan 570126, China

Abstract: To tackle semantic rigidity and insufficient cross-modal context in college English translation teaching, this study proposes an immersive multi-modal teaching strategy based on adaptive hypergraph neural networks. We develop a dual-stream hybrid hypergraph attention network (DS-HHAN) that integrates ViT and BERT for feature extraction, employs tensor fusion and dynamic gating to reduce noise, and utilises hypergraph structures to capture high-order semantic associations. Experimental results demonstrate a BLEU-4 score of 45.2 and a METEOR score of 0.83, indicating effective mitigation of semantic mismatch and enhanced contextual understanding in machine-assisted translation systems.

Keywords: English translation teaching; multimodal representation learning; dual-stream hybrid; hypergraph attention network; cross-modal tensor fusion; teaching strategy optimisation.

DOI: 10.1504/IJRIS.2026.155641

International Journal of Reasoning-based Intelligent Systems, 2026 Vol.18 No.19, pp.38 - 50

Received: 05 Mar 2026
Accepted: 26 Apr 2026
Published online: 07 Aug 2026 *