Bilingual Summarization and Readability Optimization for Textile-Industry News Using the BART Model

To address translation inconsistencies, information omission, and poor readability in bilingual summarization of textile-industry news, this study proposes a Domain-Enhanced Bidirectional and Autoregressive Transformers (DEBART) framework based on the BART architecture. A high-quality Chinese–English parallel corpus is constructed, and domain-specific terminology knowledge is incorporated into the embedding layer through a gated-fusion mechanism to improve the accuracy and consistency of professional term translation. In addition, a coverage mechanism is introduced to suppress redundant content and enhance semantic completeness, while an FKGL-driven readability optimization module dynamically adjusts sentence structures and lexical complexity to improve fluency and reader accessibility. Experimental results demonstrate that the proposed model achieves ROUGE-1, ROUGE-2, and ROUGE-L scores of 46.25, 25.76, and 42.84, respectively, outperforming conventional sequence generation methods in both summarization quality and readability. The proposed framework provides an effective solution for bilingual information processing in the textile industry and offers valuable reference for intelligent multilingual information transmission, semantic communication, and knowledge dissemination in electromagnetic-enabled communication environments and future antenna-assisted information service systems, where accurate and efficient cross-language content delivery is increasingly important.