ParroT: Translating during Chat using Large Language Models tuned with Human Translation and Feedback

TL;DR

ParroT is a framework that enhances chat-based translation capabilities of open-source large language models by fine-tuning with human translation data and feedback, improving translation quality and controllability.

Contribution

It introduces a novel instruction-based fine-tuning approach with a '$\mathbf{Hint}$' field, enabling regulation and improvement of translation performance in open-source LLMs.

Findings

Translation instruction significantly improves translation quality.

Error-guided instruction further enhances performance.

Automatic evaluation tools can effectively guide error correction.

Abstract

Large language models (LLMs) like ChatGPT have exhibited remarkable abilities on a wide range of natural language processing~(NLP) tasks, including various machine translation abilities accomplished during chat. However, these models are only accessible through restricted APIs, which creates barriers to new research and advancements in the field. Therefore, we propose ParroT, a framework to enhance and regulate the translation abilities during chat based on open-source LLMs (e.g., LLaMA), human-written translation and feedback data. Specifically, ParroT reformulates translation data into the instruction-following style, and introduces a "$\mathbf{Hint}$" field for incorporating extra requirements to regulate the translation process. Accordingly, we propose three instruction types for finetuning ParroT models, including translation instruction, contrastive instruction, and error-guided instruction. Experiments on Flores subsets and WMT22 test sets suggest that translation instruction improves the translation performance of vanilla LLMs significantly while error-guided instruction can lead to further improvement, which demonstrates the importance of learning from low-quality translations annotated by humans. We also demonstrate the potential of automatic evaluation tools in providing quality information of translations, when constructing error-guided instructions for directions that lack human annotation data. Please refer to our Github project for more implementation details: https://github.com/wxjiao/ParroT