Nongenerative Artificial Intelligence in Medicine: Advancements and Applications in Supervised and Unsupervised...

Area of Science:

• Medical Informatics
• Computational Pathology
• Artificial Intelligence in Medicine

Background:

• Artificial intelligence (AI) adoption is rapidly increasing in healthcare, transforming clinical decision support, personalized medicine, and predictive analytics.
• Current AI tools in healthcare primarily utilize nongenerative AI, specifically supervised and unsupervised machine learning (ML) techniques.

Purpose of the Study:

• To review the application of nongenerative AI methods in medicine.
• To explore how traditional AI, rooted in rules-based systems, enhances diagnostic accuracy, efficiency, and consistency.
• To discuss the performance, explainability, and reliability of these AI models for clinical decision-making.

Main Methods:

• Explanation of supervised learning models (e.g., decision trees, support vector machines, regression, K-nearest neighbor, neural networks) and their applications.
• Discussion of unsupervised learning techniques (e.g., clustering, dimensionality reduction, anomaly detection) for disease subtype discovery and outlier identification.
• Highlighting technical aspects of applying nongenerative AI algorithms to whole slide image analysis.

Main Results:

• Nongenerative AI enhances diagnostic accuracy, efficiency, and consistency in medical applications.
• Supervised and unsupervised ML techniques offer diverse capabilities from classification to novel pattern discovery.
• Analysis of whole slide images using AI shows promise for pathology.

Conclusions:

• Understanding nongenerative AI models and their limitations is imperative for safe and efficient clinical integration.
• Addressing challenges like data quality, model interpretability, and data drift is essential for reliable AI deployment.
• Nongenerative AI offers significant potential to advance medical practice and research when implemented thoughtfully.