Summary of Autocompletion Of Chief Complaints in the Electronic Health Records Using Large Language Models, by K M Sajjadul Islam et al.
Autocompletion of Chief Complaints in the Electronic Health Records using Large Language Models
by K M Sajjadul Islam, Ayesha Siddika Nipu, Praveen Madiraju, Priya Deshpande
First submitted to arxiv on: 11 Jan 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 presents a novel approach to developing machine learning models for generating Chief Complaint (CC) phrases or sentences for clinical notes. The goal is to create an autocompletion tool that suggests accurate and well-formatted phrases, alleviating the time-consuming task of documenting CCs for healthcare providers. The study uses text generation techniques, training Long Short-Term Memory (LSTM) models and fine-tuning three variants of Biomedical Generative Pretrained Transformers (BioGPT), including microsoft/biogpt, microsoft/BioGPT-Large, and microsoft/BioGPT-Large-PubMedQA. The authors also tune a prompt by incorporating exemplar CC sentences using the OpenAI API of GPT-4. The performance of the models is evaluated based on perplexity score, modified BERTScore, and cosine similarity score. BioGPT-Large exhibits superior performance compared to other models, achieving a remarkably low perplexity score of 1.65 when generating CC. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps healthcare providers by creating an autocompletion tool that suggests Chief Complaint phrases or sentences for clinical notes. This tool can save time and make it easier for doctors and nurses to document patient information. The study uses special machines called machine learning models to generate these suggestions. Three different types of models were tested: LSTM, BioGPT, and GPT-4. The best model was BioGPT-Large, which is very good at generating CC phrases. |
Keywords
* Artificial intelligence * Cosine similarity * Fine tuning * Gpt * Lstm * Machine learning * Perplexity * Prompt * Text generation
