When Professor Ming Liao started working in the United States after earning a Ph.D. in the area of Bayesian machine learning from Duke University, machine learning was still a niche area within business communities. Now, a little more than a decade later, machine learning is ubiquitous and its relevance in business is unquestioned. 

Now, behind every video shown on Douyin (TikTok), every product recommended on Tmall, every search result on Baidu, every food delivery on Meituan, machine learning is the core algorithm empowering the business operation. Companies that neglect to appreciate how machine learning helps to make better predictions and decisions to drive future growth get left behind quickly. 

In his course, Advanced Topics in Machine Learning, offered in NYU Shanghai and NYU Stern’s joint Master of Science in Data Analytics & Business Computing (MSDABC) program, Liao combines theory and hands-on practice to help students learn state-of-the-art machine learning and artificial intelligence techniques and apply them to solve real-world problems. Liao’s course covers topics such as predictive modeling with ensemble learning (xgboost / catboost), computer vision with Convolutional Neural Networks (CNN), natural language processing (NLP) with Recurrent Neural Networks (RNN) and transformer, image generation and style transfer with Generative Adversarial Networks (GAN), and Recommender system with deep learning.    

Professor Liao during the Advanced Topics in Machine Learning course

“Machine learning is no longer as far-fetched as it once sounded. The simplification of machine learning packages and ongoing built-out of pre-trained transfer learning models contributed to the democratization of machine learning,” says Liao. “Every aspiring person is able to upskill themselves in machine learning techniques, and apply them to solve business problems or explore business opportunities.” 

Liao’s course focuses on an intuitive understanding of machine learning concepts and their business applications. His approach is tailored for students with or without prior coding backgrounds alike. 

To help students conceptualize the machine learning process, Liao deploys as many metaphors as mathematical formulas. When introducing the mechanism of Convolutional Neural Network (CNN), an algorithm that has been widely used in image recognition systems, Liao compared it to how humans read letters. “Rather than looking at an entire page at once and grasping the meaning immediately, one reads word by word, line by line,” says Liao. “Similarly, when CNN classifies images of cats and dogs, it divides the whole image into boxes, and extracts particular features such as cat ears or dog noses in each unit.” 

Professor Liao advising student about the team project

Students remark on the accessibility of Liao’s teaching style. “Professor Liao has a unique way of making complex concepts fun and comprehensible,” says Guo Yating, MSDABC ‘22. “He helped us to overcome the great fear of machine learning.” 

A signature of Liao’s course are the many impactful course projects that students undertake with their newfound machine learning skills. For example, one group of students, who formerly had no prior knowledge of computer vision, developed a brain tumor radiogenomic classification algorithm to help brain cancer prognosis. The algorithm has the potential to improve the management, survival, and prospects of patients with brain cancer. 

For their efforts, the students earned a bronze medal in a kaggle competition. 

 

Another group of students developed an end-to-end machine learning based algorithmic trading system. 

A one-year backtest of their quantitative trading strategy on the JoinQuant quantitative trading platform shows significantly higher annualized return than Shanghai Stock Exchange Composite Index. 

 

Other projects during the Fall 2021 semester included social media tracking and analysis for disaster response and relief, content analysis for e-commerce, company bankruptcy prediction, flight delay prediction, Cryptocurrency prediction, Airbnb pricing, and retail store location selection.   

Advanced Topics in Machine Learning acts as a kind of boot camp for the MSDABC Capstone Projects course which Liao also directs. In the Capstone course, students work with world-class companies such as Disney, HSBC, Gucci, LVMH, Ping An, and Beike, applying the data analytics and problem-solving techniques they have gleaned from their courses to solve real-world business problems. 

Liao has extensive industry experience that inspires his teaching. Before joining NYU Shanghai, Liao held various senior positions, such as Chief Digital Officer at Huawei Consumer Business Group, General Manager of Marketing Science at Facebook, and Head of Consumer Engagement Marketing at Unilever. In these roles, Liao witnessed how powerful machine learning algorithms are helping businesses to drive extraordinary growth and achieve long-term competitive advantages. 

“Before, China was always learning from the West, but now we have created many successful examples of digital transformation,” says Liao. “The young generation should feel technologically confident and embrace the AI age.”