Characterization of AI Model Configurations for Model Reuse

Abstract

With the widespread creation of artificial intelligence (AI) models in biosciences, bio-medical researchers are reusing trained AI models from other applications. This work is motivated by the need to characterize trained AI models for reuse based on metrics derived from optimization curves captured during model training. Such AI model characterizations can aid future model accuracy refinement, inform users about model hyper-parameter sensitivity, and assist in model reuse according to multi-purpose objectives. The challenges lie in understanding relationships between trained AI models and optimization curves, defining and validating quantitative AI model metrics, and disseminating metrics with trained AI models. We approach these challenges by analyzing optimization curves generated for image segmentation and classification tasks to assist in a multi-objective reuse of AI models.