MANE 4961: Machine Learning for Engineering

Rensselaer Polytechnic Institute · Department of Mechanical, Aerospace, and Nuclear Engineering · 3 credit hours

Instructor: M Arshad Zahangir Chowdhury

Overview

Course Description: The course focuses on applications and implementation of machine learning methods. Mathematical and theoretical bases of ML models are introduced. Concepts in machine learning are used for classification and regression towards the prediction of performance characteristics of engineering systems.

Assessments: Homework, in-class assignments, in-class exams, and a final project.

Offerings: I designed and launched this course and have offered it 4 times: Spring '23, Spring '24, Spring '25, and Fall '25.

Prerequisites: MANE 2110 or permission of instructor.

Optional Texts

Aurélien Géron, Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems
Ian Goodfellow, Yoshua Bengio, and Aaron Courville, Deep Learning
Steven L. Brunton and J. Nathan Kutz, Data-Driven Science and Engineering: Machine Learning, Dynamical Systems, and Control

Student Learning Outcomes

Students who finish this course in a satisfactory manner will be able to demonstrate:

an ability to identify, formulate, and solve engineering problems using ML;
an ability to interpret ML models using multiple techniques;
an ability to handle and wrangle signals, images, and descriptive data for model building;
an ability to code, build, scale, and manage ML engineering projects;
an ability to articulate major findings of ML methods and results;
an ability to integrate ML methods into a research problem.

Topics & Schedule

The course covers four broad areas:

Basics: learning, features, ML challenges, model building, ML terms (hyperparameters, loss functions, train-test-validation, regularization), error measures, metrics, end-to-end project design, Python packages, multi-variate regression, multi-class vs. multi-label classification.
Supervised learning: kNN, MLP/ANN, logistic regression classifiers, SVM, CNNs, CART and ensemble methods.
Unsupervised learning: k-means, SVD, PCA, dimensionality reduction, data visualization.
Contemporary special topics of interest in AI/ML research as needed.

Lectures follow the sequence of topics below (subject to change).

Topic
Introduction to machine learning
Features, targets, scalers, vectors, matrices, tensors, and norms
Learning types
Learning challenges
Machine learning project design
Linear models, gradient descent
Polynomial and logistic regression, learning curves, bias-variance, regularization
Logistic regression and towards neural networks
Multinomial logistic regression, multi-class and multi-label classification, and artificial neurons
Training neural networks for classification problems; matrix algebra and backpropagation
Deep neural networks
Convolutional neural networks
Support vector machines
Unsupervised learning, singular value decomposition, principal component analysis, t-SNE
Decision trees, random forest, and ensembles
Recurrent neural networks
Special topics in AI/ML
Project discussions, presentations, and review