1
Deep Learning. Instructors: Prof. Mitesh M. Khapra and Prof. Sudarshan Iyengar, Department of Computer Science and Engineering, IIT Ropar.
FREE
This course includes
Hours of videos
3194 years, 1 month
Units & Quizzes
115
Unlimited Lifetime access
Access on mobile app
Certificate of Completion
Deep Learning has received a lot of attention over the past few years and has been employed successfully by companies like Google, Microsoft, IBM, Facebook, Twitter etc. to solve a wide range of problems in Computer Vision and Natural Language Processing. In this course we will learn about the building blocks used in these Deep Learning based solutions. Specifically, we will learn about feedforward neural networks, convolutional neural networks, recurrent neural networks and attention mechanisms. We will also look at various optimization algorithms such as Gradient Descent, Nesterov Accelerated Gradient Descent, Adam, AdaGrad and RMSProp which are used for training such deep neural networks. At the end of this course students will have knowledge of deep architectures used for solving various Vision and NLP tasks. (from nptel.ac.in)
Course Currilcum
- Lecture 01.1 – Biological Neuron Unlimited
- Lecture 01.2 – From Spring to Winter of Artificial Intelligence Unlimited
- Lecture 01.3 – The Deep Revival Unlimited
- Lecture 01.4 – From Cats to Convolutional Neural Networks Unlimited
- Lecture 01.5 – Faster, Higher, Stronger Unlimited
- Lecture 01.6 – The Curious Case of Sequences Unlimited
- Lecture 01.7 – Beating Humans at Their Own Games (Literally) Unlimited
- Lecture 01.8 – The Madness (2013-) Unlimited
- Lecture 01.9 – (Need for) Sanity Unlimited
- Lecture 02.1 – Motivation from Biological Neurons Unlimited
- Lecture 02.2 – McCulloch Pitts Neuron, Thresholding Logic Unlimited
- Lecture 02.3 – Perceptrons Unlimited
- Lecture 02.4 – Error and Error Surfaces Unlimited
- Lecture 02.5 – Perceptron Learning Algorithm Unlimited
- Lecture 02.6 – Proof of Convergence of Perceptron Learning Algorithm Unlimited
- Lecture 02.7 – Linearly Separable Boolean Functions Unlimited
- Lecture 02.8 – Representation Power of a Network of Perceptrons Unlimited
- Lecture 03.1 – Sigmoid Neuron Unlimited
- Lecture 03.2 – A Typical Supervised Machine Learning Setup Unlimited
- Lecture 03.3 – Learning Parameters: (Infessible) Guess Work Unlimited
- Lecture 03.4 – Learning Parameters: Gradient Descent Unlimited
- Lecture 03.5 – Representation Power of Multilayer Network of Sigmoid Neurons Unlimited
- Lecture 04.1 – Feedforward Neural Networks (a.k.a Multilayered Network of Neurons) Unlimited
- Lecture 04.2 – Learning Parameters of Feedforward Neural Networks (Intuition) Unlimited
- Lecture 04.3 – Output Functions and Loss Functions Unlimited
- Lecture 04.4 – Backpropagation (Intuition) Unlimited
- Lecture 04.5 – Backpropagation: Computing Gradients w.r.t. The Output Units Unlimited
- Lecture 04.6 – Backpropagation: Computing Gradients w.r.t. Hidden Units Unlimited
- Lecture 04.7 – Backpropagation: Computing Gradients w.r.t. Parameters Unlimited
- Lecture 04.8 – Backpropagation: Pseudo Code Unlimited
- Lecture 04.9 – Derivative of the Activation Function Unlimited
- Lecture 04.10 – Information Content, Entropy and Cross Entropy Unlimited
- Lecture 05.1 – Recap: Learning Parameters: Guess Work, Gradient Descent Unlimited
- Lecture 05.3 – Contours Maps Unlimited
- Lecture 05.4 – Momentum Based Gradient Descent Unlimited
- Lecture 05.5 – Nesterov Accelerated Gradient Descent Unlimited
- Lecture 05.6 – Stochastic and Mini-batch Gradient Descent Unlimited
- Lecture 05.7 – Tips for Adjusting Learning Rate and Momentum Unlimited
- Lecture 05.8 – Line Search Unlimited
- Lecture 05.9 – Gradient Descent with Adaptive Learning Rate Unlimited
- Lecture 05.9 – Bias Correction in Adam (Part 2) Unlimited
- Lecture 06.1 – Eigenvalues and Eigenvectors Unlimited
- Lecture 06.2 – Linear Algebra: Basic Definitions Unlimited
- Lecture 06.3 – Eigenvalue Decomposition Unlimited
- Lecture 06.4 – Principle Component Analysis and its Interpretations Unlimited
- Lecture 06.5 – Principle Component Analysis and its Interpretations, Part 2 Unlimited
- Lecture 06.6 – Principle Component Analysis and its Interpretations, Part 3 Unlimited
- Lecture 06.6 – Principle Component Analysis and its Interpretations, Part 3 (cont.) Unlimited
- Lecture 06.7 – Practical Example Unlimited
- Lecture 06.8 – Singular Value Decomposition Unlimited
- Lecture 07.1 – Introduction to Autoencoders Unlimited
- Lecture 07.2 – Link between PCA and Autoencoders Unlimited
- Lecture 07.3 – Regularization in Autoencoders (Motivation) Unlimited
- Lecture 07.4 – Denoising Autoencoders Unlimited
- Lecture 07.5 – Sparse Autoencoders Unlimited
- Lecture 07.6 – Contractive Autoencoders Unlimited
- Lecture 08.1 – Bias and Variance Unlimited
- Lecture 08.2 – Train Error vs Test Error Unlimited
- Lecture 08.2 – Train Error vs Test Error (Recap) Unlimited
- Lecture 08.3 – True Error and Model Complexity Unlimited
- Lecture 08.4 – L2 Regularization Unlimited
- Lecture 08.5 – Dataset Augmentation Unlimited
- Lecture 08.6 – Parameter Sharing and Tying Unlimited
- Lecture 08.7 – Adding Noise to the Inputs Unlimited
- Lecture 08.8 – Adding Noise to the Outputs Unlimited
- Lecture 08.9 – Early Stopping Unlimited
- Lecture 08.10 – Ensemble Methods Unlimited
- Lecture 08.11 – Dropout Unlimited
- Lecture 09.1 – A Quick Recap of Training Deep Neural Networks Unlimited
- Lecture 09.2 – Unsupervised Pre-training Unlimited
- Lecture 09.3 – Better Activation Functions Unlimited
- Lecture 09.4 – Better Initialization Strategies Unlimited
- Lecture 09.5 – Batch Normalization Unlimited
- Lecture 10.1 – One-hot Representations of Words Unlimited
- Lecture 10.2 – Distributed Representations of Words Unlimited
- Lecture 10.3 – SVD for Learning Word Representations Unlimited
- Lecture 10.4 – Continuous Bag of Words Model Unlimited
- Lecture 10.5 – Skip-Gram Model Unlimited
- Lecture 10.5 – Skip-Gram Model (cont.) Unlimited
- Lecture 10.6 – Contrastive Estimation Unlimited
- Lecture 10.7 – Hierarchical Softmax Unlimited
- Lecture 10.8 – GloVe Representations Unlimited
- Lecture 10.9 – Evaluating Word Representations Unlimited
- Lecture 10.10 – Relation between SVD and Word2Vec Unlimited
- Lecture 11.1 – The Convolution Operation Unlimited
- Lecture 11.3 – Convolutional Neural Networks Unlimited
- Lecture 11.3 – Convolutional Neural Networks (cont.) Unlimited
- Lecture 11.4 – CNNs (Success Stories on ImageNet) Unlimited
- Lecture 11.5 – Image Classification Continued (GoogleNet and ResNet) Unlimited
- Lecture 12.1 – Visualizing Patches which Maximally Activate a Neuron Unlimited
- Lecture 12.1 – Visualizing Patches which Maximally Activate a Neuron Unlimited
- Lecture 12.2 – Visualizing Filters of a CNN Unlimited
- Lecture 12.3 – Occlusion Experiments Unlimited
- Lecture 12.4 – Finding Influence of Input Pixels using Backpropagation Unlimited
- Lecture 12.5 – Guided Backpropagation Unlimited
- Lecture 12.6 – Optimization over Images Unlimited
- Lecture 12.7 – Create Images from Embeddings Unlimited
- Lecture 12.8 – Deep Dream Unlimited
- Lecture 12.9 – Deep Art Unlimited
- Lecture 12.10 – Fooling Deep Convolutional Neural Networks Unlimited
- Lecture 13.1 – Sequence Learning Problems Unlimited
- Lecture 13.2 – Recurrent Neural Networks Unlimited
- Lecture 13.3 – Backpropagation through Time Unlimited
- Lecture 13.4 – The Problem of Exploding and Vanishing Gradients Unlimited
- Lecture 13.5 – Some Gory Details Unlimited
- Lecture 14.1 – Selective Read, Selective Write, Selective Forget – The Whiteboard Analogy Unlimited
- Lecture 14.2 – Long Short Term Memory (LSTM) and Gated Recurrent Units (GRUs) Unlimited
- Lecture 14.3 – How LSTMs Avoid the Problem of Vanishing Gradients Unlimited
- Lecture 14.3 – How LSTMs Avoid the Problem of Vanishing Gradients (cont.) Unlimited
- Lecture 15.1 – Introduction to Encoder Decoder Models Unlimited
- Lecture 15.2 – Applications of Encoder Decoder Models Unlimited
- Lecture 15.3 – Attention Mechanism Unlimited
- Lecture 15.3 – Attention Mechanism (cont.) Unlimited
- Lecture 15.4 – Attention over Images Unlimited
- Lecture 15.5 – Hierarchical Attention Unlimited
