Unit-1 Introduction Of Neural Network
Introduction to ANN, Models of a Neuron, Activation functions, Learning and Training: Hebbian, Memory based, Competitive, Supervised and Unsupervised learning, Memory models, Recall and Adaptation, Network Architectures, Single-layered Feed-forward Networks, Multi-layered Feedforward Networks, gradient descent and contemporary variants, back-propagation algorithm, regularization, batch normalization, loss functions, Unit saturation, aka the vanishing gradient problem, and ways to mitigate it. RelU Heuristic s for avoiding bad local minima, Heuristics for faster training, Regularization, Dropout, Recurrent Networks, Topologies.
Unit-2 Convolutional Neural Networks
Convolutional Networks- Fundamentals, architectures, pooling, visualization, popular convnet architectures - AlexNet, ZFNet, VGG, C3 D, GoogLeNet, ResNet, MobileNet -v1, Inception, Training a Convnet: weights initialization, batch normalization, hyperparameter optimization. Recurrent Neural Networks LSTM, GRU, Encoder Decoder architectures.
Unit-3 Deep Unsupervised Networks
Autoencoders (standard, sparse, denoising, contractive, etc), Variational Autoencoders, Adversarial Generative Networks, and DBM.
Unit-4 Deep Learning Tools And Applications
Deep Learning Tools: TensorFlow Caffe, Theano, Torch, etc. Case study and applications in Image Processing, Natural Language Processing, Speech Recognition, Video Analytic etc. using different deep neural networks.