The Interaction Hour

Machine learning. It’s a term often used, but not always understood in the world of technology. Every day, new innovations, products and capabilities are introduced and adopted by people all over the world, but there’s a bit of a disconnect between researcher and consumer. How is a system trained? Why does it make certain decisions under certain conditions? What kind of reasoning goes into its decision making, and how can we trust that its choice is informed, objective and, ultimately, correct?

Online communities like Reddit or Twitter act like town halls, where opinions are shared and everyone, in theory, has a voice. Only, it doesn’t always work like that. What was once optimistically viewed as a solution to public discourse, offering promises of open and logical discussions where anyone with a keyboard and an internet connection could speak their piece, has instead become a bit of a Wild West. Message boards have degraded into sources of harassment, misinformation, radicalization, and more. The question is: How can you moderate, while also maintaining the promise of free speech? How can you avoid discouraging posters whose content was moderated or removed, while encouraging them to remain a part of public discourse?

We are living in a data-centric world. You might not realize, but data influences nearly every important decision you’ll make on a daily basis. Consider your daily commute from Marietta to Midtown Atlanta. You leave at a particular time or take a particular route based on your understanding of the traffic data. You choose a particular restaurant after work or select a hair stylist based on Yelp reviews. You vote and influence the entire direction of your local, national, and global community based on your understanding of political trends or voting records. There’s so much data, that an entire field – data analytics – exists to make sense of it all. But what about people like you and me? How can we, as non-data analysts, take advantage of all of this information to make decisions or come to better-informed conclusions?

Consider for a moment the story of a veteran who has returned home from a tour of duty in a combat zone in Iraq. The physical toll of war has long since worn off, but the traumatic events they witnessed or in which they participated have left mental scars that can never fully disappear. They visit mental health therapists specializing in post-traumatic stress disorder, but feel like they aren’t getting better. The anguish of reliving the experiences makes it difficult to perform the exercises their therapist has recommended, and the therapist has no clear sign of whether or not their patient is being forthcoming in each visit.

This is an imagined but common scenario for American veterans, who come home by the thousands with high rates of mental illness. Today, host Dr. Ayanna Howard is joined by Rosa Arriaga, a senior research scientist in the School of Interactive Computing, whose new grant from the National Science Foundation aims to take this challenge head-on.

What are the challenges to effective care of patients facing PTSD or other chronic illnesses? Can usable computational tools be the key to improving the effectiveness and efficiency of treatment? Why is it important that we in the computing community continue to think about how our technologies work for people in the real world?

In recent years, as computing as become central to most fields of study, so too has the education and research being performed in Georgia Tech’s College of Computing. One person who has been here through it all is Charles Isbell, the new John P. Imlay Jr. Dean of Computing. We chat with Dean Isbell about the importance of maintaining an interdisciplinary approach to research, the potential challenges facing computer science education and computing as a whole in the coming years, and why equity is the tie that binds all we do toward a fruitful future of computing.

When Zvi Galil, the outgoing John P. Imlay Jr. Dean of Computing, came to Georgia Tech in 2010, there was no such thing as OMSCS. True online degree programs were still a dream, AI teaching assistants unnecessary, and the College of Computing, while excellent, in many ways mirrored its peers in higher education. Over nearly a decade that he has helmed the College, however, it has experienced dramatic growth both in size and reputation. Due in large part to the Online Master of Science in Computer Science program that Dean Galil spearheaded, computing is now Georgia Tech’s largest major and is also a consensus top-8 computer science program nationally. As he prepares for the final month of his deanship at the College of Computing, we’ll chat today with Dean Galil about what brought him to Georgia Tech, his mission and how he fulfilled it, and, of course, the world-renowned online degree program for which he will be most remembered.

In the late 1990s, the United States saw a sharp increase in the number of opioid overdose deaths – rising by nearly 600 percent between 1999 and 2017, according to data provided by the CDC. It has, appropriately, been labeled an epidemic, and in 2018 the country’s life expectancy dropped for the third consecutive year, reflecting the ongoing drug crisis and rising suicide rates. As researchers and clinicians continue to examine the quality of different approaches to treatment, many seeking recovery have taken matters into their own hands. Our guest, School of Interactive Computing Ph.D. student Stevie Chancellor, will present a paper on this subject next week at the ACM Conference on Human Factors in Computing Systems in Glasgow, Scotland. What exactly do these addiction support communities entail? What alternative strategies are people pursuing in recovery, and why? How can we ensure that clinicians are well-informed about the types of self-treatments being used outside of their care?

In the late 1990s, Professor Gregory Abowd of Georgia Tech’s School of Interactive Computing developed a tool to allow people to collect and reflect upon memories over a long period of time. Motivated by his father’s collection of 30 years worth of videos, Gregory wanted to create something that assisted in annotating and searching videos to create short memories. Around 2002, he began using this for his own family memories and made a discovery while watching one of the videos. His oldest son, who was then 5 years old and already diagnosed with autism, demonstrated stark differences in behavior and communication between videos at 18 months and others at 26 months. Amazed by what he saw in the videos, Gregory began to consider other more serious applications of this memory-capturing tool. In the coming years, it would become a key research initiative for Gregory and others at Georgia Tech.

Over the years, virtual reality has become a mythical new medium with promises of immersive gaming and enriched experiences. Novels and movies like Ready Player One have teased the potential – and raised the expectations. In many ways, though, the technology is a largely untapped resource for reasons varying from the usability of the equipment to the premium cost.

In this episode, however, we’ll hear from former Georgia Tech student Aditya Vishwanath and current Georgia Tech assistant professor Neha Kumar who are examining the potential for virtual reality in education and instruction. What are the affordances of the technology inside of a classroom, and how can issues of cost and access be overcome to ensure it is a truly democratized medium?

Note: Inspirit is now piloting a virtual reality curriculum on fostering a civic mindset by teaching ethics and design thinking as an integrated component of a 4-year undergraduate liberal arts degree program. This project is being conducted in partnership with Krea University in Sri City, India, and will be one of the first of many VR higher-ed curricula on social justice themes at scale.

Aditya and Neha, along with colleagues Sally Creel and Tamara Pearson will speak on this topic at SXSW EDU at 5 p.m. March 4 in Room 11AB of the Austin Convention Center in Austin, Texas.

School of Interactive Computing Ph.D. student Kalesha Bullard does research into helping AI gain basic building blocks for how to learn complex tasks. In one example, she describes the goal of packing a lunch box. What are the things that a robot must know in order to complete that task? The size and shape of fruits or beverages? The heigh or circumference of each object? The depth or surface area of the lunch box itself? Taking inspiration from human learners, including her own time as a teacher and student, Bullard offers some input into how these tasks can be achieved.