Summary of Llms-based Few-shot Disease Predictions Using Ehr: a Novel Approach Combining Predictive Agent Reasoning and Critical Agent Instruction, by Hejie Cui et al.
LLMs-based Few-Shot Disease Predictions using EHR: A Novel Approach Combining Predictive Agent Reasoning and Critical Agent Instruction
by Hejie Cui, Zhuocheng Shen, Jieyu Zhang, Hui Shao, Lianhui Qin, Joyce C. Ho, Carl Yang
First submitted to arxiv on: 19 Mar 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
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 study explores the application of Large Language Models (LLMs) to convert structured patient visit data into natural language narratives, enabling zero-shot and few-shot disease prediction tasks. LLMs are trained on various EHR-prediction-oriented prompting strategies and evaluated against traditional supervised learning methods. The research proposes a novel approach using LLM agents with distinct roles: a predictor agent for making predictions and generating reasoning processes, and a critic agent that analyzes incorrect predictions to provide guidance for improving the predictor’s reasoning. The results show that the proposed approach enables decent few-shot performance in EHR-based disease prediction tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how computers can help predict diseases using patient data from electronic health records (EHRs). Usually, this requires a lot of labeled data, which can be hard to get. Researchers tried using Large Language Models (LLMs) that can understand and generate human-like text to convert structured patient data into natural language narratives. They tested different ways of prompting the LLMs and compared their results to traditional methods. The study also proposes a new approach where two LLM agents work together: one predicts diseases and generates reasons why, while the other analyzes mistakes and helps improve the predictor’s thinking. Overall, this research shows that LLMs can be useful in disease prediction tasks. |
Keywords
* Artificial intelligence * Few shot * Prompting * Supervised * Zero shot
