Deep Learning - Fall 2015

Explore fundamental and advanced concepts in deep learning through this comprehensive lecture series covering neural network architectures, optimization techniques, and practical implementations. Begin with foundational topics including perceptrons and backpropagation, then advance through regularization methods to prevent overfitting. Delve into natural language processing with Word2vec for learning word embeddings and discover probabilistic graphical models through Sum-Product Networks. Master sequential data processing with recurrent neural networks and computer vision applications using convolutional networks. Examine generative models through Restricted Boltzmann Machines and gain hands-on experience with popular deep learning frameworks including Theano and Keras. Learn from expert instruction by Ali Ghodsi alongside student presentations that demonstrate real-world applications of deep learning concepts across various domains.

Syllabus

Ali Ghodsi, Lec [1,1]: Deep Learning, Introduction
Ali Ghodsi, Lec [1,2]: Deep Learning, Perceptron, Backpropagation
Ali Ghodsi, Lec [2,1]: Deep Learning, Regularization
Ali Ghodsi, Lec [2,2]: Deep Learning, Regularization
Ali Ghodsi, Lec [3,1]: Deep Learning, Word2vec
Ali Ghodsi, Lec [3,2]: Deep Learning, Word2vec
Ali Ghodsi, Lec 11: Hard margin Support Vector Machine (svm)
Lec [4,1]: Deep Learning, Sum-Product Networks
Lec [4,2]: Deep Learning, Sum-Product Networks
Ali Ghodsi, Lec [5.1]: Deep Learning, Recurrent neural network
Ali Ghodsi, Lec [5.2]: Deep Learning, Recurrent neural network
Ali Ghodsi, Lec [6], Deep Learning: convolutional network
Ali Ghodsi, Lec [7], Deep Learning , Restricted Boltzmann Machines (RBMs)
Deep Learning Student Presentation
Deep Learning, Theano Short Tutorial
Deep Learning: Keras Short Tutorial