Design of Automatic Summarization and Translation System for English Language Documents using Long Short-Term Memory Network

Le résumé fourni par la source

In the digital age, the number of English documents continues to grow, creating an increasingly urgent need for efficient automatic summarization and translation systems. This paper proposes the design of an automatic summarization and translation system for English documents based on a long short-term memory (LSTM) network. The English documents are first preprocessed, and then the LSTM network is used to extract key information for summarization and text translation. The inclusion of a cognitive module has achieved significant results, improving the translation accuracy of culturally loaded words (such as idioms and metaphors) by 15 %, whereas without the cognitive module, translations would likely be inaccurate. The sentiment consistency of emotional text translations reached 89 %, approaching the level of human translation. The proposed system effectively improves the efficiency of English document processing.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.