This is a reading course that will investigate formalisms of "neural machines"; these include the Perceptron, the Hopfield net, Kohonen's self-organizing maps, recurrent neural nets, Valiant's neuroidal nets, echo state machines, liquid state machines, spiking neural networks, Kanerva's hyperdimensional computers, transformers, and neural Turing machines, among others.

The ultimate goal is to develop models of biological/digital systems that are able to learn from experience and adapt to changing environments. This is quite different from the Turing machine model, which is geared towards computing individual functions. To quote from Dune: "The mystery of life is not a problem to solve, but a reality to experience."

We'll begin with a few lectures that present an overview of this field. In the remaining meetings, each participant in the course will present a paper. There won't be any other requirements.

The course is intended for theory-minded students who are interested in this research area.