This course covers preliminaries of deep learning required for undergraduate students. In this course, we will start with mathematical prerequisites for deep learning and cover 3 different frameworks.  The following content will be covered in the course:

• Module 1: Mathematical foundations of Deep Learning: Scalars, Vectors, Matrices and Tensors, Multiplying Matrices and Vectors, Identity and Inverse Matrices, Linear dependence and span, Norms, Special kinds of matrices and vectors, Trace operations, Eigenvalue decomposition

• Module 2: Fundamentals of Deep Learning Anatomy of Neural Networks: Layers, Models, Loss functions and optimizers Training Deep Networks: Cost Functions, Optimizers Types of Deep Neural Networks 

• Module 3: Convolutional Neural Networks: Motivation, Convolution Operation, Types of layers, Pooling, LENET5 Architecture 

• Module 4: Recurrent Neural Networks: Architecture of traditional RNN, Types and applications of RNN, Variants of RNNs, Word Embedding using Word2vec 

• Module 5: Regularization and Autoencoders: Regularization for Deep Learning: L1 and L2, Dropout, Data Augmentation, Early Stopping, Case study on MNIST data, Autoencoders: Architecture, Implementation, Denoising Autoencoders, Sparse Autoencoders, Use cases