B.A.I. is a cutting-edge optical neural network research aiming at low-energy image classification with light. It uses optical components such as a laser, a digital micromirror device, and a spatial light modulator to encode images and perform calculations in the neural network. Relative to the traditional way of image classification which is implemented exclusively in silicon, the optical setup is able to perform computationally intensive operations and classify images from various datasets with much lower energy consumption.