Summary of Guiding Large Language Models to Post-edit Machine Translation with Error Annotations, by Dayeon Ki et al.
Guiding Large Language Models to Post-Edit Machine Translation with Error Annotations
by Dayeon Ki, Marine Carpuat
First submitted to arxiv on: 11 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Machine Translation (MT) has yet to be fully replaced by large language models (LLMs), despite their success in other areas of natural language processing. This research combines the strengths of LLMs and supervised MT systems by using external feedback, derived from Multidimensional Quality Metric (MQM) annotations, to guide LLMs in automatically post-editing MT output. The study uses LLaMA-2 models and explores different prompting strategies that provide varying levels of feedback to improve translation quality. Experimental results on Chinese-English, English-German, and English-Russian MQM data show that prompting LLMs for post-edits enhances TER, BLEU, and COMET scores. Fine-tuning the model helps integrate fine-grained feedback more effectively, leading to further improvements in translation quality. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine Translation is a big deal because it can help computers talk to each other in different languages. Right now, special computer systems are used to do this job, but large language models could also be used. This research figure out how to make these language models work better by giving them feedback on their translations. They tested this idea with three different language pairs and found that it made the translations better. The study shows that using feedback can improve translation quality, making it a useful tool for many applications. |
Keywords
» Artificial intelligence » Bleu » Fine tuning » Llama » Natural language processing » Prompting » Supervised » Translation
