DUAL-REFLECT: Enhancing Large Language Models for Reflective Translation through Dual Learning Feedback Mechanisms
This work addresses a bottleneck in enhancing translation quality for low-resource languages, representing an incremental improvement over existing self-reflection methods.
The paper tackles the problem of limited feedback in self-reflection methods for large language models in machine translation by introducing the DUAL-REFLECT framework, which uses dual learning to provide effective feedback, resulting in improved translation accuracy and ambiguity reduction, particularly for low-resource language pairs.
Recently, large language models (LLMs) enhanced by self-reflection have achieved promising performance on machine translation. The key idea is guiding LLMs to generate translation with human-like feedback. However, existing self-reflection methods lack effective feedback information, limiting the translation performance. To address this, we introduce a DUAL-REFLECT framework, leveraging the dual learning of translation tasks to provide effective feedback, thereby enhancing the models' self-reflective abilities and improving translation performance. The application of this method across various translation tasks has proven its effectiveness in improving translation accuracy and eliminating ambiguities, especially in translation tasks with low-resource language pairs.