Enabling Human-Centered Machine Translation Using Concept-Based Large Language Model Prompting and Translation Memory

This study evaluates a novel human-machine collaborative machine translation workflow, enhanced by Large Language Model features, including pre-editing instructions, interactive concept-based post-editing, and the archiving of concepts in post-editing and translation memories. By implementing GPT-4 prompts for concept-based steering in English-to-Chinese translation, we explore its effectiveness compared to traditional machine translation methods such as Google Translate, human translators, and an alternative human-machine collaboration approach that utilizes human-generated reference texts instead of concept description. Our findings suggest that while GPT-4’s discourse-level analysis and augmented instructions show potential, they do not surpass the nuanced understanding of human translators at the sentence-level. However, GPT-4 augmented concept-based interactive post-editing significantly outperforms both traditional methods and the alternative method relying on human reference translations. In testing English-to-Chinese translation concepts, GPT-4 effectively elucidates nearly all concepts, precisely identifies the relevance of concepts within source texts, and accurately translates into target texts embodying the related concepts. Nevertheless, some complex concepts require more sophisticated prompting techniques, such as Chain-of-Thought, or pre-editing strategies, like explicating linguistic patterns, to achieve optimal performance. Despite GPT-4’s capabilities, human language experts possess superior abductive reasoning capabilities. Consequently, at the present stage, humans must apply abductive reasoning to create more specific instructions and develop additional logic steps in prompts, which complicate the prompt engineering process. Eventually, enhanced abductive reasoning capabilities in large language models will bring their performance closer to human-like levels. The proposed novel approach introduces a scalable, concept-based strategy that can be applied across multiple text segments, enhancing machine translation workflow efficiency.