The Machine Learning Center at Georgia Tech (ML@GT) is home to many talented students from across campus, representing all six of Georgia Tech’s colleges and the Georgia Tech Research Institute (GTRI).
These students have diverse backgrounds and a wide variety of interests both inside and outside of the classroom. Today, we’d like you to meet Yuan Yang, a machine learning Ph.D. student who interned remotely for Amazon this summer and is focused on building reliable and controllable artificial intelligence (AI) agents that can help solve real-world problems.
Name: Yuan Yang
Hometown: Chengdu, China
Major: Machine Learning
Advisor: Le Song
Previous degrees earned and from what institutions: MS in Computational Data Science, Carnegie Mellon University and BEng in Software Engineering, Beihang University
Tell us about your research interests. Where might people be impacted by them in everyday life?
I’m interested in interpretable ML methods that learn with weak supervision. In particular, I explore models that learn by interacting with humans through dialogues and formal languages. With this human-in-the-loop learning paradigm, my research aims at building AI systems for problems where data labeling is costly, e.g. fraud detection, clinical diagnosis, and chatbots.
What drew you to wanting to research these areas?
We have seen a growing success of modern ML models in a wide range of applications, but they are also known to be data-hungry and can be difficult to train in domains where massive labels are not available. In the past, I have worked for three AI companies. Many times I found the bottlenecks that prevent us from making the AI system into a good product are not the choice of the models but the lack of labeled data and model interpretability. Therefore, I’m very interested in exploring models and methods that can learn in the “wild” and solve real-world problems.
What motivated or inspired you to pursue your Ph.D.?
I’m very fortunate to be able to witness the rise of the AI technology in the last decade during my undergraduate study. Many of the exciting achievements such as AlphaGo has deeply intrigued me and motivated me to contribute to this field by pursuing the Ph.D. in machine learning.
Earning your Ph.D. requires a lot of work. What has been challenging, rewarding or unexpected about this experience?
Doing research is hard for many reasons, and the one I find to be most challenging is the ability to “find the right problem to solve.” There are many times I have found the problems I have being trying to solve for weeks are already extensively studied in the literature. Therefore, it’s also very rewarding when I manage to solve them and find problems that I’m really excited about.
Why did you choose Georgia Tech?
Georgia Tech is the home of a group of very talented ML researchers. The ML@GT community is very active and I’m excited to work as one of them.
What’s your favorite Georgia Tech memory so far?
Seeing the Coda office built from the ground up, moving in and actually working there. The views are amazing!
What have you been up to during quarantine?
I interned at Amazon for the summer, which was a pretty interesting experience because it was my first time doing a 100% remote internship (and with 3 hour time difference) in the industry.
Though 2020 seems to have lasted a lifetime, we are still at the beginning of a new decade. What are you most looking forward to in the next 10 years?
I wish people in the world to stay strong and safe during the pandemic and hopefully getting back to our daily life soon!
What is your proudest accomplishment?
I feel proud whenever I publish a paper, and more so when I received email inquiries from people that are interested in my work.
What are some of your hobbies?
I enjoy photography, watching anime and playing video games. I’m also a big fan of classical music.
Podcast, book, movie, or tv show? Why that medium and what are some of your favorites from your chosen medium?
The Lord of The Ring trilogy is my favorite move series of all time and my recent runner-up is Interstellar.
If you could time travel to any period of time, where would you go and why?
Back to 1980s and start my own tech company 😉
What’s your favorite place to study or work and why?
I find myself being most productive in a “working” atmosphere. The thing I missed the most during this quarantine is my office in Coda.
Why do you think embodying Georgia Tech’s motto of “progress and service” is important, especially in regards to ML and AI?
As we are making progress in the technological front, AI researchers should actively think about how these technologies could impact the society both in good and bad ways. We should take the responsibility upon ourselves to make sure they will be used for social good.
mlatgt.blog 