Artificial Intelligence May Not Perform With Accurately When Analyzing Data across Multiple Healthcare Systems

Artificial Intelligence News

AI_-May_-Not_-Perform_-With_-Accurately-_When-_Analysing-_Data_-across_-Multiple_-Healthcare-_Systems Artificial Intelligence May Not Perform With Accurately When Analyzing Data across Multiple Healthcare SystemsArtificial Intelligence is the latest trend in the technology-based researches, flying high in many industries. Recent studies show that deep learning models should be tested carefully across multiple circumstances before incorporating into clinical practice.

Recently developed Artificial intelligence (AI) tools which are trained to detect pneumonia on chest X-rays, shows a serious decrease in performance when tested on data outside healthcare systems. This study was conducted at the Icahn School of Medicine at Mount where the study was published in f PLOS Medicine on Machine Learning and Health Care. This study summarizes that Artificial Intelligence (AI) mainly in the medical field must be carefully tested for its performance across populations; otherwise, these models may not perform accurately.

This research is a build on paper which was earlier published this year in the journals Radiology and Nature Medicine. The research provides the framework for applying computer vision and deep learning techniques and natural language processing algorithms, for identifying clinical concepts in radiology reports and for CT scans.