MLDNN Syllabus

Machine Learning & Deep Neural Networks (3040233305)

Pre-requisite

• Basic knowledge of Python programming, including variables, loops, and functions; understanding of data structures and numerical operations; fundamental concepts of statistics and linear algebra; and familiarity with data handling and pre-processing using libraries like NumPy and Pandas.

Course Objective

• The course introduces students to the fundamentals of Artificial Intelligence, Machine Learning, and Deep Learning, covering types of machine learning and the complete workflow from data collection to model evaluation. Students will apply supervised and unsupervised learning algorithms, understand neural network concepts, and gain hands-on experience with Python tools and libraries like NumPy, Pandas, Matplotlib, scikit-learn, TensorFlow/Keras, and PyTorch.

Teaching Scheme

Lecture	Theory	Practical	Hours	Credit
3	0	2	5	4

Unit-01 Foundations of Machine Learning (35%)

• Introduction to AI & ML, differences between ML, AI, and DL
• Types of ML: Supervised, Unsupervised, Reinforcement
• ML workflow: data collection, pre-processing, training, testing, evaluation
• Python setup, Jupyter Notebook, overview of libraries (NumPy, Pandas, Matplotlib, scikit-learn).

Unit-02 Supervised & Unsupervised Learning (28%)

• Linear regression, polynomial regression, logistic regression, KNN, SVM
• Model Evaluation: Accuracy, precision, recall, F1-score, ROC-AUC
• Unsupervised Learning: K-means clustering, hierarchical clustering, DBSCAN, PCA, LDA

Unit-03 Introduction to Deep Learning (17%)

• AI vs Machine Learning vs Deep Learning
• Applications of Deep Learning
• Basics of Artificial Neural Networks
• Perceptron model
• Activation functions: Sigmoid, Tanh, ReLU, Softmax
• Basic concepts of optimization and gradient descent

Unit-04 Deep Neural Networks (20%)

• Deep Neural Network architecture
• Forward and Backpropagation
• Loss functions (MSE, Cross-Entropy)
• Optimization techniques (SGD, Adam – overview)
• Overfitting and regularization (L1, L2, Dropout)
• Introduction to TensorFlow/Keras and PyTorch

Practical

1. Install Python and Jupyter Notebook and write a simple Python program to display a welcome message and perform basic arithmetic operations.
2. Write a Python program to import NumPy, Pandas, Matplotlib, and scikit-learn libraries and display their versions.
3. Load a CSV dataset using Pandas and display the first five rows, last five rows, and basic information of the dataset.
4. Visualize a dataset using Matplotlib by plotting a bar chart, line chart, and histogram.
5. Implement Simple Linear Regression using scikit-learn on a given dataset.
6. Implement Polynomial Regression and compare its results with Linear Regression
7. Implement Logistic Regression for a binary classification problem
8. Implement the K-Nearest Neighbors (KNN) algorithm for classification and analyze the effect of different values of K
9. Evaluate a classification model by calculating Accuracy, Precision, Recall, and F1-Score.
10. Plot the ROC curve and calculate the AUC score for a binary classification model.
11. Apply K-Means Clustering on a dataset and visualize the formed clusters.
12. Perform Hierarchical Clustering on a dataset and plot the dendrogram
13. Apply Principal Component Analysis (PCA) for dimensionality reduction on a dataset.
14. Implement a simple Perceptron model or basic Neural Network using scikit-learn or Keras and demonstrate training and prediction.

---