Summary of Is Larger Always Better? Evaluating and Prompting Large Language Models For Non-generative Medical Tasks, by Yinghao Zhu et al.
Is larger always better? Evaluating and prompting large language models for non-generative medical tasks
by Yinghao Zhu, Junyi Gao, Zixiang Wang, Weibin Liao, Xiaochen Zheng, Lifang Liang, Yasha Wang, Chengwei Pan, Ewen M. Harrison, Liantao Ma
First submitted to arxiv on: 26 Jul 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); 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 A study benchmarks various Large Language Models (LLMs), including GPT-based LLMs, BERT-based models, and traditional clinical predictive models, for non-generative medical tasks using renowned datasets. The study assesses 14 language models and 7 traditional predictive models on the MIMIC dataset and TJH dataset, focusing on tasks such as mortality and readmission prediction, disease hierarchy reconstruction, and biomedical sentence matching. Results show that LLMs exhibit robust zero-shot predictive capabilities on structured EHR data using well-designed prompting strategies, often outperforming traditional models. However, for unstructured medical texts, LLMs do not outperform finetuned BERT models, which excel in both supervised and unsupervised tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A study compares different types of computer programs called Large Language Models (LLMs) to see how well they can help doctors make decisions using patient data. The researchers tested 14 LLMs and 7 other kinds of computer programs on two big datasets, each containing medical records from hospitals. They used the computer programs for tasks like predicting when patients might die or come back to the hospital, identifying diseases, and matching medical descriptions. The study found that some LLMs can make good predictions about structured patient data without being trained on it first, which is useful. However, other LLMs don’t do as well with unstructured patient notes, like doctor’s notes. This shows that different computer programs are better for different tasks and types of medical information. |
Keywords
» Artificial intelligence » Bert » Gpt » Prompting » Supervised » Unsupervised » Zero shot
