AMATH 649 - Neural Networks

An introduction to neural network methods, with some discussion of their relevance to neuroscience. Simple neuron models and networks of neurons. Training feedforward networks for classification or regression. Learning using the backpropagation of errors. Unsupervised learning methods. Optimal linear decoding. Recurrent neural networks. Convolutional neural networks. Advanced topics, including adversarial inputs and biologically plausible learning methods.