Introduction to machine learning-
Linear models (SVMs and Perceptron’s, logistic regression)- Intro to Neural Nets: What a shallow network computes- Training a network: loss functions, back propagation and stochastic gradient descent- Neural networks as universal function approximates
DEEP NETWORKS:
History of Deep Learning- A Probabilistic Theory of Deep LearningBackpropagation and regularization, batch normalization- VC Dimension and Neural Nets-Deep Vs Shallow NetworksConvolutional Networks- Generative Adversarial Networks (GAN), Semisupervised Learning
DIMENTIONALITY REDUCTION:
Linear (PCA, LDA) and manifolds, metric learning - Auto encoders and dimensionality reduction in networks - Introduction to Convnet - Architectures – AlexNet, VGG, Inception, ResNet - Training a Convnet: weights initialization, batch normalization, hyperparameter optimization
OPTIMIZATION AND GENERALIZATION
Optimization in deep learning– Non-convex optimization for deep networks- Stochastic Optimization Generalization in neural networksSpatial Transformer Networks- Recurrent networks, LSTM - Recurrent Neural Network Language Models- Word-Level RNNs & Deep Reinforcement Learning - Computational & Artificial Neuroscience
CASE STUDY AND APPLICATIONS
Imagenet- Detection-Audio Wave Net-Natural Language Processing Word2Vec - Joint Detection BioInformatics- Face Recognition- Scene UnderstandingGathering Image Captions
Question Paper Pattern:
• The Question paper will have TEN questions
• Each full question will be for 20 marks
• There will be 02 full questions (with maximum of four sub questions) from each module.
• Each full question will have sub questions covering all the topics under a module.
• The students will have to answer FIVE full questions, selecting one full question from each module.
Textbooks :
1. CosmaRohillaShalizi, Advanced Data Analysis from an Elementary Point of View,2015.
References:
1. Deng & Yu, Deep Learning: Methods and Applications, Now Publishers,2013.
2. Ian Goodfellow, YoshuaBengio, Aaron Courville, Deep Learning, MIT Press,2016. Michael Nielsen, Neural Networks and Deep Learning, Determination Press,2015.
Course Outcomes:
1. Demonstrate the basics of deep learning for a given context.
2. Implement various deep learningmodels for the given problem
3. Realign high dimensional data using reductiontechniques for the given problem
4. Analyze optimization and generalization techniques of deeplearning for the given problem.
5. Evaluate the given deep learningapplication and enhance by applying latest techniques.