A Bayesian machine based on memristors

Over the past few decades, the performance of machine learning models on various real-world tasks has improved significantly. Training and implementing most of these models, however, still requires vast amounts of energy and computational power.