Summary of Large Language Models Struggle in Token-level Clinical Named Entity Recognition, by Qiuhao Lu et al.
Large Language Models Struggle in Token-Level Clinical Named Entity Recognition
by Qiuhao Lu, Rui Li, Andrew Wen, Jinlian Wang, Liwei Wang, Hongfang Liu
First submitted to arxiv on: 30 Jun 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 The paper explores the potential of Large Language Models (LLMs) in token-level Named Entity Recognition (NER) for clinical texts, particularly in the context of rare diseases. Current research focuses on document-level NER, whereas this study investigates the effectiveness of both proprietary and local LLMs in token-level clinical NER using techniques like zero-shot prompting, few-shot prompting, retrieval-augmented generation (RAG), and instruction-fine-tuning. The authors’ experiments reveal the challenges faced by LLMs in token-level NER for rare diseases and propose potential improvements for their application in healthcare. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The study looks at how Large Language Models (LLMs) can be used to help doctors find important information in medical texts. Currently, most research focuses on finding specific words or phrases across an entire document, rather than pinpointing where the information is located within that text. This paper wants to see if LLMs can do better by using these models for “token-level” NER, which involves identifying the location of important information within a text. The researchers tested different ways of using LLMs and found some challenges, but also some potential solutions. |
Keywords
» Artificial intelligence » Few shot » Fine tuning » Named entity recognition » Ner » Prompting » Rag » Retrieval augmented generation » Token » Zero shot
