Unit$-$I: Learning System and evaluation parameters$[4$ Hours$]$

Well posed learning problem, Designing a learning system, Issues in AI learning. Hypothesis Space, Hypothesis functions, Hypothesis Evaluation, Bias, Variance, Underfitting, Overfitting, Inductive bias, Evaluation, Training, Testing, $$Cross$-$validation. Error Analysis, Error Metrics, Precision and recall.

Unit II: Introduction to Artificial Neural Networks $($ANNs$)[5$ Hours$]$

Neural Computation, History of Artificial Neural Systems Development, Biological Neurons and Their Artificial Models, Models of Artificial Neural Networks, Neural Network Learning Rules, Perceptron Learning Algorithms. Supervised learning, unsupervised learning, and reinforcement learning, Training Process: Overview of the neural network training process and the role of training data., Initialization of Weights: Importance of initializing weights and common strategies $($e$.$g$.,$ random, Xavier, He initialization$).$ Hebbian Learning, Competitive, Boltzmann Learning,

Unit III: Single Layer Perceptron Classifiers$[5$ Hours$]$

Classification Model, Features, and Decision Regions, Discriminant Functions, Linear Machine and Minimum Distance Classification, Nonparametric Training Concept, Training and Classification Using the Discrete Perceptron: Algorithm and Example, Single$-$Layer Continuous Perceptron Networks for Linearly Separable Classifications, Multicategory Single$-$Layer Perceptron Network $$

Unit IV: Multilayer Feedforward Networks$[5$ Hours$]$

Linearly Nonseparable Pattern Classification, Delta Learning Rule for Multi$-$Perceptron Layer, Generalized Delta Learning Rule, Feedforward Recall and Error Back$-$Propagation Training, Learning Factors, Classifying and Expert Layered Network, Multilayer Networks, Backpropagation algorithm, case study to implement MLP$.$

UNIT V Competitive learning Neural Network $[5$ Hours$]$

Components of CL network, Pattern clustering and feature mapping network, ART networks, Features of ART models, character recognition using ART network. Self$-$Organization Maps, $($SOM$)$: Two Basic Feature Mapping Models, Self$-$Organization Map, SOM Algorithm, Properties of Feature Map, Computer Simulations, Learning Vector Quantization, Adaptive Pattern Classification. Application and analysis of ART$1$ & ART$2,$ Case study

Unit VI: Optimization of Neural Networks$[4$ Hours$]$

Data Preparation and Preprocessing, Weight Initialization Techniques, Loss Functions $($Mean Squared Error $($MSE$)$ and Cross$-$Entropy Loss.$)$ and Backpropagation, Learning Rate strategies and its Importance, Gradient Descent Variants $($SGD$,$ Adam, RMSprop$),$ Overfitting and Regularization Techniques $($L$1,$ L$2,$ Dropout$),$ Hyper$-$parameter Tuning, Handling Vanishing and Exploding Gradients