These notes were developed for CISC 5800: Machine Learning at Fordham University during Spring 2024. They provide a mathematically grounded introduction to core machine learning concepts, with an emphasis on derivations, intuition, and algorithmic understanding.
While originally created to support classroom instruction, these notes were significantly expanded beyond what was presented in lectures. In many cases, additional derivations, examples, and explanations are included to provide a more complete and self-contained reference.
The notes are designed for students with a background in calculus, linear algebra, and basic probability. The focus is on understanding why machine learning algorithms work—not just how to use them.
Topics include:
- Linear Regression and Logistic Regression
- Gradient Descent and Optimization
- Neural Networks and Backpropagation
- Support Vector Machines (including dual formulation)
- Clustering Methods (k-Means, Hierarchical, DBSCAN, Fuzzy Clustering)
- Expectation Maximization and Gaussian Mixture Models
- Introductory discussion of modern topics
These notes aim to:
- Develop a mathematical foundation for machine learning
- Connect theory to algorithms through explicit derivations
- Provide geometric and conceptual intuition where possible
- Serve as a standalone reference for students and instructors
- The material goes well beyond the pace of a typical lecture sequence
- Many derivations and explanations are included that were not covered in class
- The goal was to create a resource that students could revisit after the course
The full set of notes is available as a PDF:
📄 ML_Lecture_Notes.pdf
These notes occupy a middle ground between:
- purely conceptual introductions to machine learning
- and highly abstract, research-level treatments
They are intended to be accessible but rigorous, particularly for students in mathematics, data science, or related fields.
These notes were developed for instructional purposes and reflect the structure and emphasis of CISC 5800 at the time they were taught. They are not intended to be a comprehensive treatment of all machine learning topics.
Brent Young, Ph.D. Associate Professor of Mathematics Wilkes University