Summary of Automated Clinical Data Extraction with Knowledge Conditioned Llms, by Diya Li et al.
Automated Clinical Data Extraction with Knowledge Conditioned LLMs
by Diya Li, Asim Kadav, Aijing Gao, Rui Li, Richard Bourgon
First submitted to arxiv on: 26 Jun 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 novel framework aligns generated internal knowledge with external knowledge through in-context learning (ICL), effectively increasing the accuracy of extracting lung lesion information from clinical and medical imaging reports. Large language models (LLMs) are employed to interpret unstructured text, but often hallucinate due to a lack of domain-specific knowledge. The framework uses a retriever to identify relevant units of internal or external knowledge and a grader to evaluate truthfulness and helpfulness, updating the knowledge bases. Experimental results on expert-curated test datasets demonstrate an average increase in F1 score for key fields (lesion size, margin, and solidity) by 12.9% over existing ICL methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to help computers better understand medical reports is being developed. These reports often contain important information about lung diseases, but it’s hard for computers to extract that information accurately because they don’t always have the right knowledge. A team has created a new framework that uses large language models and other tools to help computers learn from both their own internal knowledge and external sources of information. This approach can improve the accuracy of extracting important details like lesion size, margin, and solidity by an average of 12.9% compared to current methods. |
Keywords
» Artificial intelligence » F1 score
