AMATH 449 - Neural Networks

An introduction to neural network methods, with some discussion of their relation to neuroscience. Simple neuron models and networks of neurons. Training feedforward networks. Backpropagation of errors and stochastic gradient descent. Unsupervised learning methods. Recurrent neural networks. Convolutional neural networks. Continuous time models. Adversarial attacks. Advanced topics may include neural engineering, biologically plausible learning methods.