NUS Symposium on Artificial Intelligence Education

Monday 21 February 2022

9.30 am to 4 pm

Abstracts

AI Education in a Fast-Moving World

(Keynote Panel Discussion)

In the past decade, AI is reborn with explosive new opportunities in the research community and more importantly, in our daily lives. AI technologies have ushered in in new eras in natural language understanding, image recognition, autonomous driving, etc. They have also brought new challenges in, e.g., privacy, fairness, and transparency, in the age of massive data. The fundamental challenge of AI remains: shed light on intelligence, human or not. The opportunities ahead abound: empower the solutions to many social, economic, and environmental challenges in the post-pandemic world. How do we best teach AI in universities going forward? How do prepare our students to face the opportunities and challenges ahead?

The topics of interest include but not limiting to:

• What are the fundamental theories and tools that form the core curriculum of an AI education at the undergraduate and graduate level?
• What are the main challenges and opportunities in teaching different AI courses with the rapid advancements of the field with new research results and methods?
• How to balance theory and practice in the different AI modules?
• What are some of the effective methods of teaching AI in a multidisciplinary settings?

 Moderator: Prof. Lee Wee Sun, Head, Department of Computer Science, NUS School of Computing.

Panel Members:

Civic AI Education: Embedding Civic Values in Learning Computer Science

A/Prof. Zhang Weiyu, Department of Communications and New Media

Artificial Intelligence offers tremendous opportunities but raises concerns over harming civic values. AI Education has focused on advancing technologies but lagged in embedding civic values in the learning. Civic AI Education highlights the procedure leading to the AI outcomes, by centering on the participation of diverse citizens who embrace civic values in the entire process of AI design, implementation and monitoring. The topics covered in Civic AI Education include, but are not limited to, data privacy, data security, transparent AI, explainable AI, AI fairness (or biases), and AI auditing.

When AI meets Healthcare: Special Considerations and Challenges

Asst. Prof. Feng Mengling, NUS Saw Swee Hock School of Public Health

In the recent boom of AI technologies, we have witness many successful application of AI solutions in various industries ranging from finance, logistics, robotics, security and autonomous vehicle. But we have not seen as many success in deploying AI solutions for solving challenges in healthcare though. This is NOT a random co-incident. In this talk, I will share a couple of case studies based on the real experiences from projects of my Healthcare AI lab to highlight some unique challenges and considerations that we should be mindful of when developing AI solutions for healthcare problems.

Lessons Learned in Teaching Coding to Non-CS NUS Students and Developing MOOCs that Reach Thousands

Asst. Prof. Clayton, Department of the Built Environment

In this talk, I plan to share strategies from five years of teaching basic coding and data science concepts to students from the Project and Facilities Management course in the Dept. of the Built Environment. These strategies include the use of subjective group objectives rather than individual coding memorization to take advantage of the diversity of interest and rate of coding skills development amongst a small group of students. The goal is to set some of the students on a path towards data science, while exposing other students to coding, but not expecting full proficiency. I will also discuss the development of an international award-nominated Massive Open Online Course (MOOC) that I developed for NUS that has been on the edX platform since April 2020 and has had over 24,000 participants. This course is called Data Science for Construction, Architecture, and Engineering and focuses on the context-based exposure and development of Python skills for this industry.

Essential Software Engineering skills for Artificial Intelligence and Data Science

Dr. Ganesh Neelakanta Iyer, Department of Computer Science

While dealing with AI and DS based projects, there are a few key software engineering skills needed. It includes choosing the right platform, applying Engineering Best Practices (EBP) such as testing and continuous integration for those projects. For example, MLOps is a set of practices that aims to deploy and maintain machine learning models in production reliably and efficiently. Students (from Non-programming background specifically) needs to understand these essential principles so that they can develop and deploy their projects more efficiently, reliably and in a scalable manner.

Vignettes from Teaching the Theory of Machine Learning to Mathematics Majors

A/Prof. Vincent Tan, Department of Electrical and Computer Engineering and Department of Mathematics

In this spiel, I’ll talk about my multi-year experiences teaching MA4270, whose clientele includes undergraduates from the Department of Mathematics and cognate departments. Different from most other ML/DS courses on campus, here we focus on the *theory* of machine learning, and students are expected to understand and do proofs. I will attempt to convince the audience that course serves an important purpose, especially for students who wish to understand why certain methods work and when they fail, and not simply on how to use readily available packages.

Key Issues in Designing Artificial Intelligence, Data Science and Analytics Curricula and Modules

(Town Hall Discussion)

Artificial Intelligence, Data Science and Analytics stand at the crossroads of programming, technology and mathematics and creating curricula and modules in these areas present unique challenges. Hear from our distinguished panel about their experiences designing and teaching such modules to a wide spectrum of NUS students in various disciplines, and seek their opinions on any issues that you may be facing in your own modules.

 Moderator: Dr. Daren Ler, Lecturer, Department of Computer Science, NUS School of Computing.

Panel Members: