Supervised and Unsupervised Learning

Course Objectives :                            Why this Subject?

1. To learn about supervised learning and its algorithms
2. To learn about supervised learning networks
3. To learn about unsupervised learning
4. To learn about Unsupervised learning network, Autoencoder, Generative Adversarial Network

Course Outcomes (CO)

• CO1 Applying classification and regression algorithm to real world examples
• CO2 Learn about supervised learning network
• CO3 Applying clustering algorithms to real world examples
• CO4 Learn about Unsupervised learning network, Autoencoder, Generative Adversarial Network

UNIT-I

Supervised Learning: Classification, Ensemble Learning Classification-Introduction, Example of Supervised Learning, Classification Model, Classification Learning Steps, Common Classification Algorithms - k-Nearest Neighbour (kNN), Decision tree, Random forest model, Support vector machines. Ensemble Learning- Boosting, Bagging

Regression: Introduction, Regression Algorithms - Simple linear regression, Multiple linear regression, Polynomial Regression Model, Logistic Regression

Bayesian Concept Learning - Introduction, Bayes’ Theorem, Bayesian Belief Network.

UNIT-II

Supervised Learning Networks:

Perceptron: Representational power of Perceptron, The Perceptron Training Rule, Gradient Descent and Delta Rule

Multilayer Networks: A differentiable Threshold Unit, Representational Power of Feedforward Networks

Backpropagation Algorithm: Convergence and local minima, Hypothesis space search and Inductive Bias, Generalization, overfitting and stopping criteria.

Regularization for Deep Learning: Parameter Norm Penalties, Dataset Augmentation, Noise Robustness, Early Stooping, Sparse Representation, Dropout.

Optimization for Training Deep Models: Challenges in Neural network Optimization, Basic Algorithms, Parameter Initialization Strategies.

UNIT-III

Unsupervised Learning: Introduction, Unsupervised vs Supervised Learning, Application of Unsupervised Learning, Clustering –Clustering as a Machine Learning task, Different types of clustering techniques, Partitioning methods, Hierarchical clustering, Density-based methods: DBSCAN.

Finding Pattern using Association Rule: Definition of common terms, Association rule, Apriori algorithm.

UNIT – IV

Unsupervised Learning Networks: Kohonen Self-Organizing Feature Maps – architecture, training algorithm, Kohonen Self Organizing Motor Map,Restricted Boltzmann machine

Autoencoders: Linear Factor Methods such as Probabilistic PCA and Factor Analysis, Independent Component Analysis, Sparse Coding; Undercomplete Autoencoders, Regularized Autoencoders, Stochastic Encoders and Decoders, Denoising Autoencoders, Contractive Autoencoders, Applications of Autoencoders.

Generative Adversarial Networks: Generative Vs Discriminative Modeling, Probabilistic Generative Model, Generative Adversarial Networks (GAN), GAN challenges: Oscillation Loss, Mode Collapse, Uninformative Loss, Hyperparameters, Tackling GAN challenges, Wasserstein GAN, Cycle GAN, Neural Style Transfer

Textbook(s):

1. Tom M. Mitchell, “Machine Learning”, McGraw-Hill Education (India) Private Limited, 2013 
2. Jennifer Grange ,” Machine Learning for Absolute Beginners: A Simple, Concise & Complete Introduction to Supervised and Unsupervised Learning Algorithms”,Kindle 
3. Kamal Kant Hiran, Ritesh Kumar Jain, Dr. Kamlesh Lakhwani, Dr Ruchi Doshi,” Master Supervised and Unsupervised Learning Algorithms with Real Examples”, BPB Publications

References:

1. C. M. BISHOP (2006), “Pattern Recognition and Machine Learning”, Springer-Verlag New York, 1st Edition 
2. Michael W. Berry, Azlinah Mohamed,Bee Wah Yap,”Supervised and Unsupervised Learning for Data Science”