De Kai is a Professor of Computer Science and Engineering at HKUST, a Distinguished Research Scholar at Berkeley’s International Computer Science Institute, and the CTO of Ahura AI. He is among only 17 scientists worldwide named by the Association for Computational Linguistics as a Founding ACL Fellow, for his pioneering contributions to machine translation and machine learning foundations of systems like the Google/Yahoo/Microsoft translators. He completed his PhD at UC Berkeley and holds an executive MBA from Kellogg (Northwestern University) and HKUST. His work on machine learning of human languages and the cognitive relationships between them has produced over 120 scientific papers. Amidst the many different titles he has held and the countless awards and accolades he’s received over time, De Kai is, at his core, a lifelong learner. 

In his words, “I’ve always been motivated by really wanting to accomplish some task.” This concept is at the heart of De Kai’s approach to education as both a learner and a teacher. “I think it’s really important to understand what the objective is and to have that be an interesting objective.” Purpose and practical application, according to De Kai, are necessary prerequisites to meaningful learning in the classroom. With decades of experience as both a student and a teacher in various capacities, De Kai has developed a sense for which types of classes and activities tend to stick with students over time. He’s found that, both for himself and for his students, interactive hands-on work is the most effective way to foster true understanding and discovery. It starts with setting a goal or objective and progresses to interactive work that leads students through a process of guided struggling with the goal of solving an open problem. It’s in that process of struggle and cognitive effort that true learning and understanding take place. De Kai recalls the most impactful classes from his undergraduate and doctoral studies as the ones which, at the time, got the lowest evaluations from students. They were the classes that were so difficult and had such intense workloads that students were forced to work together just to get through it all. He described the classes as being, “highly interactive in the sense that you were engaging and struggling interactively with either the material or the instructor. And because those tasks were interesting and relevant years later, they are still, in many cases, the only thing that I remember.”

De Kai strives to structure the classes he teaches in a similar way. He teaches one class, “…where in the course of one semester, [he takes] the students through all of C and C++ through every different programming paradigm.” Through a series of assignments that build up to a project, students learn how to build a programming language by actually building their own programming language. Needless to say, it’s an extremely difficult course that requires extensive amounts of time and effort to complete. In his words, “It’s tough, but boy, do they understand programming and programming languages after that.” While many students may resent the work in the moment, they always walk away with a comprehensive and practical understanding of the material. He explains that, “…what learners think is the right thing to do in the moment is not…necessarily what ends up being the most useful…Part of the work of an instructor is to be able to anticipate where they are going to find things useful and then find a way for them to engage with that…the philosophy is to focus on…not just ‘Are you having fun right now?’ but ‘Are you going to feel like this really paid off?’”.

De Kai’s lifelong study of what works in the classroom and what doesn’t has influenced his approach and decision to work with Ahura AI. He explains Ahura like this: teaching, “…is not the same as just hitting play on a YouTube video. What you do when you’re actually addressing a classroom or a lecture is you’re looking at the audience. You’re looking at the learners, and you’re interacting with them. You’re making eye contact. You’re sensing: Are they getting it? Are they confused? Are they bored or off doing something else? And you adjust how you’re delivering, how you’re teaching, and how you’re interacting.” We often lose this sense of interaction and adjustment with digital content. As a teacher, once you’ve identified the goals for learning, De Kai says that it becomes a question of, “…how do you find the most efficient way to keep the learner engaged? And so I think the goal of Ahura is to use AI to do that.” De Kai went on to say, “We’re in an era where AI is making huge progress technologically. Why shouldn’t that be applied to one of the most important things in our economy and people’s lives…why shouldn’t we be using AI to help people learn effectively? That’s the goal in all the things Ahura is aiming to do.”

Do you agree with De Kai’s approach to teaching? Have you ever taken a class that was so difficult that students were forced to work together just to get through it? Which teaching methods have worked the best for you over time? Let us know your answers to any of these questions and your thoughts on this weeks #HowILearn spotlight in the comments or in a #HowILearn post of your own. Be sure to tag @ahuraai and we’ll be sure to get back to you!

Author: Alex Murray, Operations and Client Success Intern, UC Berkeley ‘23