COMPSCI 589: Machine Learning (Grade: A)

Course Overview:
This course provided a comprehensive introduction to core machine learning models and algorithms, covering key concepts in classification, regression, clustering, and dimensionality reduction. It balanced theoretical understanding with practical application, allowing students to grasp both the mathematical foundations and the real-world usage of machine learning techniques.

Course Structure and Content

Theory Side:
The course delved into the mathematical underpinnings of widely-used machine learning algorithms. Topics included optimization methods (both global and local), linear regression, logistic regression, support vector machines (SVM), and kernel methods. A strong emphasis was placed on understanding the relationships between different models and algorithms, fostering a deep comprehension of the concepts.

Applied Side:
On the applied front, the course focused on leveraging machine learning techniques to solve real-world problems. Key areas included model selection, regularization, experimental design, and the presentation and interpretation of results. The course’s assignments required students to not only solve mathematical problems but also implement algorithms, providing a hands-on experience that bridged theory with practice.

Class Meetings

• In-person Sections:
  • Tuesday and Thursday, 5:30pm-6:45pm (Engineering Lab II Room 119)
• Online Section:
  • ECHO360 recordings available on Canvas.

Instructor

The course was instructed by Hui Guan, who brought a wealth of knowledge and expertise to the classroom, guiding students through the intricate details of machine learning.

Course Management Tools

• Canvas for course management.
• CampusWire for Q&A.
• Gradescope for assignment grading.

Course Schedule Highlights

The course was spread over 14 weeks, with each week focusing on a specific topic or set of topics:

• Week 1-3: Introduction to Optimization
  Covered global and local optimization techniques, with a focus on gradient descent.

• Week 4-5: Linear Models
  Explored linear regression, logistic regression, and support vector machines (SVM).

• Week 6-8: Unsupervised Learning
  Introduced Principal Component Analysis (PCA) and K-means clustering, followed by feature engineering techniques like normalization and feature selection.

• Week 9-10: Nonlinear Learning and Feature Learning
  Discussed nonlinear learning approaches, feature learning, and universal approximators.

• Week 11-14: Advanced Topics
  Covered kernel methods, Multi-Layer Perceptron (MLP), decision trees, and midterm recap. The course concluded with a comprehensive project that allowed students to apply their knowledge to a real-world problem.

Assignments and Projects

The course assignments were a blend of theoretical exercises and practical implementation tasks. Each assignment built upon the previous one, helping students to progressively master complex concepts. The final project was a culmination of everything learned, requiring a thorough application of machine learning methods to a significant problem.

Personal Experience

I successfully completed this course with an A Grade (4.0/4.0), reflecting my strong understanding of both the theoretical and practical aspects of machine learning. The combination of rigorous coursework, insightful lectures, and hands-on assignments made this one of the most enriching learning experiences in my academic journey.

Final Thoughts

This machine learning course provided me with a solid foundation in the field, equipping me with the skills needed to tackle complex data-driven challenges. The balance of theory and application was particularly valuable, as it ensured that I not only understood the underlying concepts but also knew how to apply them effectively in real-world scenarios.