Exploring quantum probability interpretations through artificial intelligence

Abstract

The varying interpretations of quantum probability governing quantum measurements are significantly reflected in divergent opinions on the foundations of statistics, including ensemble-frequency theory, propensity theory, and subjective degrees of reasonable belief. Although it has been suggested that a series of progressively sophisticated tests using artificial intelligence could yield increasingly significant experimental data to constrain potential resolutions to the measurement problem, no feasible experimental designs have yet been proposed. In this work, we utilize advanced deep learning technology to develop a novel experimental framework that integrates neural network-based artificial intelligence into a Bell test. This framework challenges the implicit assumptions underlying Bell tests. We demonstrate our framework through a simulation and introduce three new metric-morphing polygons, averaged Shannon entropy, and probability density map-to analyze the results. This approach enables us to determine whether quantum probability aligns with any one of these three interpretations or a hybrid of them.