A field of AI that deals with the interaction between computers and human (natural) languages. A neural network architecture is often used in computer vision use cases that has inductive biases around the spatial importance of the data. Basically, it assumes that things near each other are more related than things far away.