Topic: machine-learning
Deep Reinforcement Learning
Preface In classical reinforcement learning, you learnt about the agent : environment interaction, Markov Decision Process, dynamic…
Model Free Methods
Preface Monte-Carlo Methods * Points to consider Monte-Carlo Prediction Monte-Carlo Control Off Policy Importance Sampling Temporal…
Model Based Methods - Dynamic Programming
Preface Dynamic Programming Policy Iteration Steps Algorithm Policy Evaluation - Prediction * Points to consider Policy Improvement…
Fundamental Equations of Reinforcement Learning
RL Equations - State Value Function Expected value Questions RL Equations - Action Value Function Example Questions Understanding the RL…
Classical Reinforcement Learning
Preface The Evolution of RL What is Reinforcement Learning Questions Agent-Environment Interaction Example Two Types of Tasks Questions…
Modifications to Neural Networks
Purpose Deep Learning is a subset of machine learning and hence it borrows a lot of common use cases from it. Here we will be looking at the…
Backpropagation
Neural Network Topology Training a Neural Network Mathematically Complexity of the Loss Function Example Gradient Descent Questions Gradient…
Convolutional Neural Networks (CNNs)
Purpose CNN - Introduction Challenges in Image Processing CNNs - A specialised architecture for visual data Applications of CNNs Receptive…
Long, Short-term Memory, Gated Recurrent Unit
Bi-Directional RNN Problems with Vanilla RNNs Long, Short-term Memory Networks Characteristics of an LSTM Cell LSTM Cell Common Activation…
Recurrent Neural Networks (RNNs)
Recurrent Neural Networks What are Sequences RNN Formulation Architecture of RNN The flow of information in RNNs is as follows Example…
Convolutional Neural Networks (CNNs) - Industry Applications
Neural networks in industry applications Data Preprocessing: Shape, Size and Form Images - Channels and sizes Question Images…
Gaussian Mixture Models
Purpose Limitations of K-Means Advantages of GMMs over K-Means Limitations of Fitting the Data using a Single Gaussian 1D Gaussian vs…
Support Vector Machines
Support Vector Machines Maximal Marginal Classifier Soft Margin Classifier Summary Questions Kernels Boundary Transformation Feature…
Principal Component Analysis
Principal Component Analysis Problem with having a lot of features Applications of PCA Dimensionality Reduction PCA Definition PCA - Change…
Model Selection - Practical Considerations
Business Case Identifying the consumers of an e-commerce website Based on Gender Based on Age Benefits of Identifying the right consumer…
Unsupervised Learning
Types of Learning Unsupervised Learning Practical applications of Clustering Clustering vs Segmentation Segmentation Stability Inter and…
Boosting
Purpose Bagging vs Boosting Bagging Boosting Question Weak Learners Question AdaBoost Steps Question Amount of Say Change of Weight…
Ensembles and Random Forest
Purpose Ensembles Idea Diversity Ways to achieve diversity Acceptability Why ensembles work Example Coin Toss Explanation Voting Stacking…
Decision Trees
Purpose About Decision Trees Building Decision Trees Top Down Greedy Question Hyperparmaeters Hyperparameter Tuning Difference between…
Logistic Regression in the Industry
Purpose of Logistic Regression Difference between Linear Regression and Logistic Regression Why is Logistic Regression so common Confusion…