相关实验视频

Updated: Jun 28, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

大型语言模型有助于生成电子健康记录表现型算法

Chao Yan¹, Henry H Ong¹, Monika E Grabowska¹

¹Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN 37203, United States.

Journal of the American Medical Informatics Association : JAMIA

|April 13, 2024

概括

大型语言模型 (LLM) 可以起草电子健康记录 (EHR) 类型化算法. 虽然GPT-4和GPT-3.5显示出有希望的结果,但仍需要专家审查来改进.

科学领域:

医疗信息学医疗信息学
人工智能在医学中的应用
观察性健康研究观察性健康研究

背景情况:

电子健康记录 (EHR) 对观察性健康研究至关重要,但表型化需要广泛的专家投入,限制了可扩展性.
开发准确的表型算法是一个耗时的过程,涉及文献审查和证据合成.

研究的目的:

调查大型语言模型 (LLM) 在生成高效和高质量的EHR表型算法草稿方面的潜力.
评估不同LLM在创建可执行的表型算法中的性能.

主要方法:

四名LLM (GPT-4,GPT-3.5,Claude 2,Bard) 被要求生成基于SQL的表型算法,用于2型糖尿病,痴呆症和甲状腺功能低下症.
算法遵循一个共同的数据模型 (CDM),并由三个表型专家对指令遵循,逻辑和可执行性等指标进行评估.
实施了顶级算法,并与来自eMERGE网络的现有临床医生验证的算法进行了比较.

主要成果:

与Claude 2和Bard相比,GPT-4和GPT-3.5在指令遵循,逻辑和SQL可执行性方面获得了更高的专家评估分数.
虽然LLM识别了相关的临床概念,但他们在逻辑上组织标准方面遇到了困难,导致算法过于限制或过于广泛.
这导致了召回 (限制性) 或积极预测值 (广泛) 的问题.

结论:

通过确定与CDM一致的相关临床标准,GPT-3.5和GPT-4可以生成表型化算法草案.
在信息学和临床经验方面的重大专业知识对于评估和完善LLM生成的算法至关重要.
虽然LLM显示出提高EHR表型化效率的潜力,但人类监督至关重要.

关键词:

聊天GPT 聊天GPT 聊天电子健康记录是电子健康记录.大型语言模型.现型定制现型定制现型定制

更多相关视频

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

In Vivo Modeling of the Morbid Human Genome using Danio rerio

In Vivo Modeling of the Morbid Human Genome using Danio rerio

Published on: August 24, 2013

相关概念视频

Background and Environment Affect Phenotype

Background and Environment Affect Phenotype

Although the genetic makeup of an organism plays a major role in determining the phenotype, there are also several environmental factors, such as temperature, oxygen availability, presence of mutagens, that can alter an organism’s phenotype.
An example of how genetic background affects phenotype can be seen in horses. The Extension gene in horses is responsible for their coat color. A wild-type gene (EE) produces black pigment in the coat, while a mutant gene (ee) produces red pigment. A...

Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving

Mechanistic Models: Compartment Models in Algorithms for Numerical Problem Solving

Mechanistic models play a crucial role in algorithms for numerical problem-solving, particularly in nonlinear mixed effects modeling (NMEM). These models aim to minimize specific objective functions by evaluating various parameter estimates, leading to the development of systematic algorithms. In some cases, linearization techniques approximate the model using linear equations.
In individual population analyses, different algorithms are employed, such as Cauchy's method, which uses a...

Polygenic Traits

Polygenic Traits

When more than one gene is responsible for a given phenotype, the trait is considered polygenic. Human height is a polygenic trait. Studies have uncovered hundreds of loci that influence height, and there are believed to be many more. Due to the high number of genes involved, as well as environmental and nutritional factors, height varies significantly within a given population. The distribution of height forms a bell-shaped curve, with relatively few individuals in the population at the...

Genome-wide Association Studies-GWAS

Genome-wide Association Studies-GWAS

Genome-wide association studies or GWAS are used to identify whether common SNPs are associated with certain diseases. Suppose specific SNPs are more frequently observed in individuals with a particular disease than those without the disease. In that case, those SNPs are said to be associated with the disease. Chi-square analysis is performed to check the probability of the allele likely to be associated with the disease.
GWAS does not require the identification of the target gene involved in...