Scholarlink

A method for English translation error recognition combining reinforcement learning and evolutionary computation was studied, which utilizes 3D ResNet to extract translation action features and Transformer to generate Japanese text, improving online accuracy.The multi population mechanism is used to increase the diversity of the population. Each subpopulation selects a specific solution strategy through the adaptive mechanism, and then each subpopulation evolves separately. The Sigmoid activation function processes translation features, evaluates error rates, and identifies and outputs errors if they are high. The experiment proves that this method has high accuracy and effectively identifies targets.

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 Automatic English Online Translation Error Recognition Method Based on Reinforcement Learning and Evolutionary Computation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Xia Li

摘要

A method for English translation error recognition combining reinforcement learning and evolutionary computation was studied, which utilizes 3D ResNet to extract translation action features and Transformer to generate Japanese text, improving online accuracy.The multi population mechanism is used to increase the diversity of the population. Each subpopulation selects a specific solution strategy through the adaptive mechanism, and then each subpopulation evolves separately. The Sigmoid activation function processes translation features, evaluates error rates, and identifies and outputs errors if they are high. The experiment proves that this method has high accuracy and effectively identifies targets.