Schedule:

2 sessions each 1 hour on every Saturday (10:00 AM -12 PM) in the month of April

Topics:

Week 1 -1st April

(Session 1) Course Introduction - Deep Learning, applications, current state of art, future direction and research

(Session 2) Neural Networks Demystified: ANN, Higher-level representations, image features, Optimization, stochastic gradient descent, Backpropogation)

Week 2 -8th April

Training Neural Networks 1 and 2

Part 1 (Session 1): activation functions, weight initialization, gradient flow, batch normalization

babysitting the learning process, hyperparameter optimization

Part 2 (Session 2): parameter updates, ensembles, dropout

Week 3 -15th April

(Session 1) Convolution Neural Net (CNN ), and case studies Alexnet, Googlelenet, VGGNet, ResNet etc.

(Session 2) CNN for Spatial Localization/ Object Detection, Visualizing CNN layers & Adverserial Examples

Week 4 -22nd April

(Session 1) Recurrent Neural Nets (RNN) and

(Session 2) Long Short Term Memories (LSTM) image captioning & Deep Reinforcement Learning (if time permits)

LAB Work

Weekly: Assignments/Lab work in Python with TesnorFlow or Theano DL libs to be finalized- HW requirements Deep Learning Workstation (must)