Summary of Context-aware and Style-related Incremental Decoding Framework For Discourse-level Literary Translation, by Yuanchang Luo et al.
Context-aware and Style-related Incremental Decoding framework for Discourse-Level Literary Translation
by Yuanchang Luo, Jiaxin Guo, Daimeng Wei, Hengchao Shang, Zongyao Li, Zhanglin Wu, Zhiqiang Rao, Shaojun Li, Jinlong Yang, Hao Yang
First submitted to arxiv on: 25 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 report presents a novel approach to translating literary texts from Chinese to English, focusing on the nuances of idiomatic expressions, narrative structures, and contextual meanings. To address these challenges, we leveraged the Chinese-Llama2 model, enhanced through continual pre-training and supervised fine-tuning. Our methodology includes an incremental decoding framework that ensures sentence-level translations consider broader context, maintaining coherence and consistency throughout. This approach captures long-range dependencies and stylistic elements, producing translations that preserve original literary quality. Experiments demonstrate significant improvements in BLEU scores for both sentence-level and document-level translations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about finding a way to translate Chinese books into English in a way that keeps the meaning and style of the original text. The problem with translating literature is that it often uses special expressions, complex storytelling, and subtle meanings that are hard to capture. To solve this challenge, we used a special machine learning model called Chinese-Llama2, which was trained to get better at this task by practicing on lots of texts. We also developed a new way of decoding sentences so that they fit together smoothly like a puzzle. This approach helps the model understand long-range connections and stylistic elements in the text, resulting in translations that are as good as the original book. The results show that our method is much better than other approaches at translating literary texts. |
Keywords
» Artificial intelligence » Bleu » Fine tuning » Machine learning » Supervised
