Summary of Tcm-ftp: Fine-tuning Large Language Models For Herbal Prescription Prediction, by Xingzhi Zhou et al.
TCM-FTP: Fine-Tuning Large Language Models for Herbal Prescription Prediction
by Xingzhi Zhou, Xin Dong, Chunhao Li, Yuning Bai, Yulong Xu, Ka Chun Cheung, Simon See, Xinpeng Song, Runshun Zhang, Xuezhong Zhou, Nevin L. Zhang
First submitted to arxiv on: 15 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE)
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 approach aims to predict Traditional Chinese Medicine (TCM) prescriptions by leveraging pre-trained large language models (LLMs) with supervised fine-tuning on a novel dataset, DigestDS. This dataset comprises practical medical records from experts in digestive system diseases. The method, TCM-FTP, incorporates data augmentation and a low-rank adaptation technique to enhance computational efficiency. Experimental results show that TCM-FTP achieves an F1-score of 0.8031, outperforming previous methods. Moreover, it demonstrates remarkable accuracy in dosage prediction. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about using computers to help create traditional Chinese medicine prescriptions. These prescriptions are combinations of herbs that have been used for thousands of years to treat different health problems. The problem is that we don’t have a lot of data on which to base our predictions, and the relationships between symptoms and herbs can be complex. To solve this problem, the authors created a new dataset called DigestDS, which contains information about digestive system diseases from experienced doctors. They then used this dataset to fine-tune pre-trained language models to make more accurate predictions. The results show that their method is much better than previous approaches at predicting TCM prescriptions. |
Keywords
» Artificial intelligence » Data augmentation » F1 score » Fine tuning » Low rank adaptation » Supervised
