What you’ll learn in a Machine Learning course

Machine Learning is a rapidly growing field that can transform many industries. It requires talented people who can leverage off-the-shelf toolboxes to implement a function while understanding the underlying mechanisms. The UIC Online Master of Engineer with a focus area in AI and Machine Learning program will equip students with exactly that background.

Meet Dr. Xinhua Zhang, UIC’s Associate Professor in the Department of Computer Science for the College of Engineering. Prior to joining UIC in 2015, Dr. Zhang was a NICTA-endorsed PhD student in the Research School of Computer Science at the Australian National University.

Dr. Zhang teaches CS 412: Intro to Machine Learning at UIC. In his course, he recognizes the disruptive technologies that can emerge suddenly and surprise even some of the greatest minds. His course prepares students for these surprises and equips them with the rapidly changing skillset sought by the job market.

Learn more about the CS 412 course, and the importance of balancing theory and practice.

What will Master of Engineering (MEng) students learn in CS 412: Intro to Machine Learning (ML)? Why is it important for engineering students to learn about ML? 

This course introduces common machine learning tasks (e.g., classification, density estimation, clustering) and some of the successful machine learning techniques and broader paradigms that have been developed for these tasks. The course is programming-intensive and a large emphasis will be placed on tying existing machine learning techniques to specific real-world applications through hands-on experience.

By taking the course, students will be able to explain the mechanism of different common machine learning task settings and be able to apply existing machine learning tools to a broad range of data sets. They will also be able to identify appropriate problem formulations and machine learning techniques for framing and solving predictive data analysis problems, along with evaluating and measuring the performance of different learning techniques on predictive tasks.

Are there any projects or real-world applications that students can look forward to in the CS 412 course? 

The course features eight Python programming projects based on real datasets, one for each week. These projects help students to code up learning algorithms from scratch. For most students, this will be the first time they’ve put learned math into practice, with a focus on scalability, extensibility, and an end-to-end pipeline from data processing up to final result visualization. Projects include shallow neural networks, decision trees, cross validation, k-nearest neighbor, naïve Bayes classifier, and data embeddings and visualization.

How and why did you become interested in machine learning? 

I have been interested in ML ever since my undergraduate thesis, where I applied fuzzy-neural networks to network routing. The area was fascinating because it keeps a good balance between practice and theory – one can develop an algorithm that delivers superior prediction or training results, while at the same time one can also prove statistical theorems to justify why it works so well. Such a balance is diminishing in recent years, because the practice of deep learning is already far ahead of its theoretical understanding. This brings a lot of opportunities for machine learning research.

What is the future for AI/ML in industry and workforce development? How will the MEng program prepare/equip students for the current workplace climate they are facing? 

While artificial general intelligence used to be considered far away, it is now becoming more and more realistic. Generative AI will continue to make inroads to a vast number of fields that impact real lives.

Disruptive technologies do not loom gradually; they emerge suddenly and surprise even some of the greatest minds (e.g., ChatGPT, AlphaGo). While it is difficult to predict what specific sub-area of ML will gain the highest popularity, the AI/ML industry will always eagerly seek talents who can not only proficiently leverage off-the-shelf toolboxes to implement a function, but also systematically understand their underlying mechanisms including mathematics. This skill cannot be garnered by simply participating in bootcamps or browsing piecemeal online blogs/tutorials. The MEng program equips students with exactly such a background, preparing them for the rapidly changing skillset sought by the job market.

What advice do you have for your students interested in pursuing their MEng degree? Why should they choose UIC? 

My advice is to keep a balance between theory and practice. Do not downplay one of them to favor the other. It may bring a lot of immediate pleasure by building up small-scale learning models based on some existing code base. However, when one delves into more complicated systems, theory will start to provide more and more valuable guidelines, insights, and prescience because simple trial-and-error would no longer be feasible.

UIC faculty provides expertise in both theory and practice. They are conducting cutting-edge research while also incorporating the latest research into their teachings. The MEng program has experts from multiple data science domains with extensive teaching experience. It prepares students with a well-balanced and carefully designed pathway to the final degree.