Polo Chau Finds Inspiration at the Intersection of Usability and Data Mining

School of CSE's Chau earns promotion to associate professor and tenure 

“If people don’t know how to use something, they won’t buy it,” said Polo Chau, associate director for Georgia Tech’s MS in Analytics program and associate professor in the School of Computational Science and Engineering (CSE).

“When I was in college in Hong Kong, I learned that computing is not narrowly defined. Now, we know about usability, which is a critical part of virtually all technologies that we use,” he continued.

Chau was recently promoted to associate professor in CSE and awarded tenure. His research bridges data mining and human-computer interaction (HCI) to synthesize scalable, interactive tools that help people understand and interact with graph datasets and machine learning models, with a focus on cybersecurity applications.

Tailoring his interests

The need for this crossover of fields comes from the age of big data itself. As massive datasets are now common in all fields, Chau focuses on the need to learn how to make sense of the data and how to visualize the findings. This is how data mining with HCI has come to be an underlying part of everything from cybersecurity and analytics to interactive computing.

The progression to this interdisciplinary research focus was one Chau has tailored to his own interests, after first becoming interested in online fraud detection while studying information engineering at the Chinese University of Hong Kong. While attending the program in the early 2000s, he found that design and utility were still not weighted equally. Computing classes often graded students on the correctness of systems and tools, but not on their usability.

This discovery only increased his resolve to find and pursue a master’s degree that incorporated his two passions in the way he envisioned they could be equally utilized.

“I had to ask what encompassed design and computing together. I searched the web and learned that human-computer interaction programs were my best match, and I applied. I got an offer from Carnegie Mellon, and this is when I began a lot of the research which is related to what I do today,” he said.

Finding the bad guys

While attending Carnegie Mellon University (CMU) for an M.S. in Human-Computer Interaction, Chau’s first research was in fraud detection for eBay, or as he puts it, “finding the bad guys.”

“I started the early part of the research for my undergraduate senior year project in Hong Kong, identifying suspicious transaction patterns among some buyers and sellers on the auction site, but I could not figure out the technique to find the bad guys. Then, when I went to Carnegie, I thought that my chances of finding someone who knew how to find these bad guys was good.”

“I wrote three emails and only got one response: from Professor Christos Faloutsos. I wrote to him asking if he would like to do an independent study as I had found an interesting problem. It was coincidental, because he wrote back immediately saying, ‘Yes!’ and he asked me to come to his office.”

As it happened, Faloutsos was working on graph algorithms for analyzing large network data, the type of research that paired well with the eBay problem, which addresses the problem of uncovering identities on the site that are likely controlled by the same fraudster.

Heading to Georgia Tech

“This was the very first paper I submitted and it was for a workshop, and he stayed up until 3 a.m. with me working on it. I thought, ‘Wow!’ I was a 20-something-year-old and I was amazed that he stayed up with me, as I knew it wasn’t normal or typical. He was the most cheerful and most positive person I have ever known and this impacted me,” he said.

After completing his doctorate in machine learning at CMU, Chau joined Georgia Tech in 2012 as an assistant professor in CSE and as an adjunct assistant professor in the School of Interactive Computing (IC).

“Georgia Tech was an easy choice because it is very collaborative and it has an open door policy. Faculty don’t lock themselves in their offices, which is important to me as my work is interdisciplinary. It is essential for me to work with colleagues and students across domains.”

Since his time at Georgia Tech, Chau maintains the accessibility and open door policy he experienced as a student. This approach to teaching has earned him the admiration and respect of his students and colleagues.

Chau continues to collaborate with many schools and units at Georgia Tech, including the School of Computer Science, IC, and the Georgia Tech Visualization Lab. His research group, the Polo Club of Data Science, has been working on a wide range of projects, from visualization for ML and deep learning, large graph mining and visualization, to developing systems to protect the security of ML-based systems.