Summary of Towards Chapter-to-chapter Context-aware Literary Translation Via Large Language Models, by Linghao Jin et al.
Towards Chapter-to-Chapter Context-Aware Literary Translation via Large Language Models
by Linghao Jin, Li An, Xuezhe Ma
First submitted to arxiv on: 12 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG)
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 In this paper, researchers address two key challenges in machine translation: discourse phenomena in document-level datasets are sparse, and most existing methods rely on unrealistic sentence-level alignments. To overcome these issues, the authors create a novel dataset of Chinese-English literature with intricate discourse structures and propose a more practical context-aware translation setting called chapter-to-chapter (Ch2Ch) translation. They investigate the performance of various machine translation models under this setting and introduce an approach to fine-tuning large language models (LLMs) for Ch2Ch literary translation, achieving significant improvements over baselines. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about how machines can be taught to translate books from one language to another in a way that takes into account the story being told. Right now, machine translation isn’t very good at this because it doesn’t understand how sentences fit together to tell a story. The authors of this paper create a new dataset of Chinese and English book translations and test different ways that machines can be taught to do this kind of translation. They find that some methods work much better than others and that using really powerful language models can help improve the quality of the translations. |
Keywords
» Artificial intelligence » Discourse » Fine tuning » Translation
