Summary of Qibo: a Large Language Model For Traditional Chinese Medicine, by Heyi Zhang and Xin Wang and Zhaopeng Meng and Zhe Chen and Pengwei Zhuang and Yongzhe Jia and Dawei Xu and Wenbin Guo
Qibo: A Large Language Model for Traditional Chinese Medicine
by Heyi Zhang, Xin Wang, Zhaopeng Meng, Zhe Chen, Pengwei Zhuang, Yongzhe Jia, Dawei Xu, Wenbin Guo
First submitted to arxiv on: 24 Mar 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 The proposed two-stage training approach combines continuous pre-training and supervised fine-tuning for Large Language Models (LLMs) in traditional Chinese medicine (TCM). The approach addresses challenges such as theory-practice gaps, lack of corpus resources, and overconfident predictions. A notable contribution is the processing of a 2GB TCM corpus, constructing pre-training and instruction fine-tuning datasets. Qibo-Benchmark evaluates LLM performance on multiple dimensions, including subjective, objective, and three TCM NLP tasks. The trained model, named Qibo, exhibits significant performance boosts compared to baselines. The average subjective win rate is 63%, objective accuracy improved by 23% to 58%, and Rouge-L scores for the three TCM NLP tasks are 0.72, 0.61, and 0.55. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In traditional Chinese medicine (TCM), a new approach to training Large Language Models (LLMs) is needed. The current method of fine-tuning LLMs may not work well because it doesn’t take into account the unique characteristics of TCM. To solve this problem, researchers developed a two-step process that includes pre-training and fine-tuning. They also created a large dataset for TCM and built a tool to test how well the model performs. The results show that their approach is much better than before, with an average win rate of 63% on subjective tasks and improved accuracy on objective tasks. |
Keywords
» Artificial intelligence » Fine tuning » Nlp » Rouge » Supervised
