Scholarlink

<p>The increased demand for precise and contextual English translation in real-time communication prompts the combination of Deep Learning (DL) and Internet of Things (IoT) systems, as traditional translation systems often overlook contextual and multimodal aspects, which limit their flexibility and competence in dynamic and heterogeneous communication scenarios. We proposed research to create an English Translation Enhancement Technology (ETET) that improves the accuracy of translation, its fluency, and contextual relevance with the help of IoT-based multimodal data and DL algorithms. The ETET architecture makes use of IoT devices such as smart watches and cellphones, and wearable sensors to fetch real-time text, voice, and context information. The preprocessing data stages include tokenization, normalization, and word2vec embedding. Bidirectional Encoder Representations of Transformers and Attention-Based Memory Network (BERT-ATTMN) model are found within the main translation system. The core applied is the translation core, which employs BERT to extract context, ATTMN to extract coherence that depends on memory, and the Namib Beetle Optimization (NBO) algorithm to optimize parameters. Python experimentation demonstrates that ETET yields better results in the Bilingual Evaluation Understudy (BLEU) score of 0.96, a precision of 97.67%, and high fluency and low latency, which are superior compared to the existing models. The contextual comprehension and translation efficiency of ETET in real-time IoT settings is effective, which shows a high possibility of its application in education, healthcare, multilingual conferencing, and international communication.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An intelligent English translation system fusing IOT contextual inputs with optimized deep learning models

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月24日

查看全文
  • Wei Zhao

摘要

The increased demand for precise and contextual English translation in real-time communication prompts the combination of Deep Learning (DL) and Internet of Things (IoT) systems, as traditional translation systems often overlook contextual and multimodal aspects, which limit their flexibility and competence in dynamic and heterogeneous communication scenarios. We proposed research to create an English Translation Enhancement Technology (ETET) that improves the accuracy of translation, its fluency, and contextual relevance with the help of IoT-based multimodal data and DL algorithms. The ETET architecture makes use of IoT devices such as smart watches and cellphones, and wearable sensors to fetch real-time text, voice, and context information. The preprocessing data stages include tokenization, normalization, and word2vec embedding. Bidirectional Encoder Representations of Transformers and Attention-Based Memory Network (BERT-ATTMN) model are found within the main translation system. The core applied is the translation core, which employs BERT to extract context, ATTMN to extract coherence that depends on memory, and the Namib Beetle Optimization (NBO) algorithm to optimize parameters. Python experimentation demonstrates that ETET yields better results in the Bilingual Evaluation Understudy (BLEU) score of 0.96, a precision of 97.67%, and high fluency and low latency, which are superior compared to the existing models. The contextual comprehension and translation efficiency of ETET in real-time IoT settings is effective, which shows a high possibility of its application in education, healthcare, multilingual conferencing, and international communication.