New paper: The Power of Combining Data and Knowledge: GPT-4o is an Effective

[maykcaldas]

Paper: The Power of Combining Data and Knowledge: GPT-4o is an Effective

Authors: Danqing Hu, Bing Liu, Xiaofeng Zhu, Nan Wu

Abstract: Lymph node metastasis (LNM) is a crucial factor in determining the initialtreatment for patients with lung cancer, yet accurate preoperative diagnosis ofLNM remains challenging. Recently, large language models (LLMs) have garneredsignificant attention due to their remarkable text generation capabilities.Leveraging the extensive medical knowledge learned from vast corpora, LLMs canestimate probabilities for clinical problems, though their performance hashistorically been inferior to data-driven machine learning models. In thispaper, we propose a novel ensemble method that combines the medical knowledgeacquired by LLMs with the latent patterns identified by machine learning modelsto enhance LNM prediction performance. Initially, we developed machine learningmodels using patient data. We then designed a prompt template to integrate thepatient data with the predicted probability from the machine learning model.Subsequently, we instructed GPT-4o, the most advanced LLM developed by OpenAI,to estimate the likelihood of LNM based on patient data and then adjust theestimate using the machine learning output. Finally, we collected three outputsfrom the GPT-4o using the same prompt and ensembled these results as the finalprediction. Using the proposed method, our models achieved an AUC value of0.765 and an AP value of 0.415 for LNM prediction, significantly improvingpredictive performance compared to baseline machine learning models. Theexperimental results indicate that GPT-4o can effectively leverage its medicalknowledge and the probabilities predicted by machine learning models to achievemore accurate LNM predictions. These findings demonstrate that LLMs can performwell in clinical risk prediction tasks, offering a new paradigm for integratingmedical knowledge and patient data in clinical predictions.

Link: https://arxiv.org/abs/2407.17900

Reasoning: Reasoning: Let's think step by step in order to determine if the paper is about a language model. We start by examining the title and abstract. The title mentions "GPT-4o," which is a large language model developed by OpenAI. The abstract discusses the use of this LLM in combination with machine learning models to improve the prediction of lymph node metastasis in lung cancer patients. The focus is on leveraging the medical knowledge embedded in the LLM and integrating it with data-driven machine learning models. This clearly indicates that the paper involves the application and capabilities of a language model.