Summary of Choose the Final Translation From Nmt and Llm Hypotheses Using Mbr Decoding: Hw-tsc’s Submission to the Wmt24 General Mt Shared Task, by Zhanglin Wu et al.
Choose the Final Translation from NMT and LLM hypotheses Using MBR Decoding: HW-TSC’s Submission to the WMT24 General MT Shared Task
by Zhanglin Wu, Daimeng Wei, Zongyao Li, Hengchao Shang, Jiaxin Guo, Shaojun Li, Zhiqiang Rao, Yuanchang Luo, Ning Xie, Hao Yang
First submitted to arxiv on: 23 Sep 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 This paper submits a machine translation model to the WMT24 shared task, specifically translating English into Chinese. The team uses various training strategies like regularized dropout and bidirectional training to develop their neural machine translation (NMT) model based on the deep Transformer-big architecture. Additionally, they employ continue pre-training, supervised fine-tuning, and contrastive preference optimization for large language models (LLMs). By applying Minimum Bayesian risk (MBR) decoding to select the best translation from multiple hypotheses, their submissions achieve competitive results in the evaluation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a machine that can translate English into Chinese. The team who made this model uses special tricks to make it better, like giving it more training and making sure it’s good at both directions of translation. They also use a big model to help with this process. The final step is to choose the best translation from all the options they got. This submission did pretty well in the competition. |
Keywords
» Artificial intelligence » Dropout » Fine tuning » Optimization » Supervised » Transformer » Translation
