Summary of Probabilistic Medical Predictions Of Large Language Models, by Bowen Gu et al.
Probabilistic Medical Predictions of Large Language Models
by Bowen Gu, Rishi J. Desai, Kueiyu Joshua Lin, Jie Yang
First submitted to arxiv on: 21 Aug 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 The paper explores the limitations of Large Language Models (LLMs) in producing reliable prediction probabilities, which are essential for transparency and decision-making in clinical applications. The authors investigate explicit prompts that generate probability estimates, but find that these underperform compared to implicit probabilities derived from the likelihood of predicting correct label tokens. They demonstrate this discrepancy across six advanced LLMs and five medical datasets, highlighting the need for improved methods and cautious interpretation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper is about how big language models are not good at making predictions with confidence levels. These models are used in medicine to make predictions, but they can’t really give a number that says how sure they are. The researchers tried using special prompts to get the model to give a probability, but it didn’t work well. They found that if you use a different way of looking at the data, you can get better results. This is important because in medicine, we need to be able to trust the predictions. |
Keywords
* Artificial intelligence * Likelihood * Probability
