"SMARTICLES” are simple, basic robots, developed by physics professor Daniel Goldman and used in a variety of experiments at Georgia Tech. Put one of them on a table and it’ll wave its arms around. Alone, it can’t do much more than that.
But put a group of them together on the table and something interesting happens.
“They start bumping against each other, and you see this thing start moving, so you’re getting this sort of unexpected locomotion,” explains Dana Randall, ADVANCE Professor of Computing. “We are working to model this algorithmically, and we’re seeing that the movement comes from this jostling between the smarticles, where you don’t get it from just the one, but by putting them together, you’re getting much more interesting, much more complex behavior.”
Randall and her colleagues in the School of Physics at Arizona State University are using the small machines to study swarm robotics. However the experiment is an apt metaphor for something much larger and demonstrates that, beyond what can be attained in isolation, collaboration often delivers unexpected and remarkable results.
In other words, it’s a lot like GT Computing.
A Vibrant Research Ecosystem
The College of Computing is the nexus of a vibrant and thriving research ecosystem. Comprising three schools, 150-plus associated research institutes, centers, labs, and groups, the GT Computing research community is populated by scores of faculty and student researchers.
These researchers work at a variety of levels, sharing new ideas and new insights with collaborators from the Georgia Tech campus – as well as from government, academia, and industry – to deliver innovative, data-driven solutions for real-world challenges.
During the past few years, GT Computing researchers have been leading or supporting novel explorations in artificial intelligence, big data analysis, cybersecurity, graph analytics, high performance computing, machine learning, robotics, and more. Ultimately, much of this groundbreaking work will go toward improving health care, education, business, community, and other people-intensive systems.
What drives this research and the flourishing community behind it is what sets it apart: a truly interdisciplinary approach and an unwavering commitment to fostering a productive research environment.
Breaking Down Traditional Silos
“Collaboration and interdisciplinary work are truly emphasized and rewarded at Georgia Tech. Not all institutions fully appreciate the real value of this approach. However, really novel and transformative results can be achieved by looking at the same problem from completely different perspectives,” Randall says.
This emphasis on collaboration begins with Georgia Tech’s interdisciplinary research institutes (IRIs).
“The IRI structure is breaking down traditional silos and bringing people together to simultaneously approach research with lenses from many different disciplines, which is frankly needed at this point,” says Randall, who co-directs Georgia Tech’s newest IRI, the Institute for Data Engineering and Science (IDEaS), along with Professor Srinivas Aluru of the School of CSE.
“No aspect of science is going to continue the way it was. New ways of analyzing and integrating the massive datasets being generated in most fields of science and engineering are transforming those disciplines and challenging long-held notions about what can be accomplished.”
IDEaS is one of 12 IRIs at Georgia Tech. It, along with three others – the Institute for Information Security and Privacy (IISP), the Institute for People and Technology (IPaT), and the Institute for Robotics and Intelligent Machines (IRIM) – are closely affiliated with the College of Computing, either through their leadership or the number of participating College faculty.
“We have a very robust presence in the IRIs, which is stronger than one might think given the College’s size relative to engineering and the sciences. GT Computing faculty leads iDEaS, IISP, and IPaT. And, up until last year, IRIM was also led by a computing faculty member,” says Ron Arkin, who in June 2017 stepped down after nine years as the College of Computing’s associate dean for research.
The IRIs support Georgia Tech’s core research areas. They are designed to be gateways for public and private academic, research, and corporate organizations. The IRIs also serve as central points of contact for key funding agencies such as the National Science Foundation, the Defense Advanced Research Projects Agency, and the National Institutes of Health.
Beyond this structure however, Arkin says the College wants to ensure that researchers who do not directly align with a particular institute or center can still feel like there is a place for them and their work at GT Computing.
“It’s not all about IRIs, it’s about creating a highly productive research environment for all of our faculty and students. It’s about enabling connections and engaging all the voices instead of a cabal going off into a corner and saying, ‘this is what we are going to do,’” Arkin says.
For many researchers, these connections are happening in dozens of centers, labs, and groups that are associated with the GT Computing-affiliated IRIs.
These units often reside within the College’s three schools. While also highly interdisciplinary in their approach, these units are more singularly focused than the IRIs. They explore more of the nuts-and-bolts aspects of computational science, which frequently serve as stepping-stones for broader, more advanced research initiatives.
Interdisciplinary research centers (IRCs) support affiliated IRIs and bring together student, faculty, and scientist researchers from other Georgia Tech units to focus on common research questions. Some of the more high-profile IRCs that are part of the GT Computing research ecosystem include the Center for Machine Learning, the GVU Center, and the Center for Research into Novel Computing Hierarchies.
Perhaps not so well known, but vitally important to a broad spectrum of research initiatives, is the Algorithms and Randomness Center (ARC).
“Algorithms are fundamental to everything happening on campus,” says Randall. “If you really want to understand something scientifically or really want a principled approach, you need to be talking to the algorithms people.”
Professors associated with IRCs and other centers like the ARC, often sponsor research labs and groups. Here, graduate and undergraduate students work hand-in-hand with professors on projects directly related to professors’ particular research interests. These groups expose students to a variety of research opportunities, helping them to define their own interests as they advance in their studies.
Resources for All Students
Also helping them to sharpen their focus and advance their careers, the College offers a number of research-oriented Ph.D. programs. These include computer science, robotics, computational science and engineering, and human-centered computing. This year, a machine learning Ph.D. program was approved to begin accepting students this fall. It is a collaborative effort between eight schools at Georgia Tech.“One of the things Georgia Tech does very well is finding ways to empower students to integrate knowledge across multiple fields,” says Randall. “Because of Georgia Tech’s interdisciplinary approach, students get vertical and horizontal training: Faculty are training their students, but the students work across collaborative groups that are focusing on single research questions.”
“One of the unique advantages that GT Computing has is that it consists of three distinct schools, which leads to many collaborative opportunities,” says Computational Science and Engineering Ph.D. candidate Bo Xie.
"This is exemplified in machine learning, which touches many disciplines. There are many top-notch faculty members who are leaders in each of these areas, and it presents abundant opportunities for collaboration and working on new research directions,” explains Xie, who specializes in training neural networks.
Undergraduate students are also a part of the research community at GT Computing. The Undergraduate Research Opportunitiesin Computing (UROC) is the College’s mechanism for pairing research-interested undergrads with GT Computing faculty.
“Research allows students to interact oneon-one with faculty and go deeper into subjects than they could in the classroom. For students considering Ph.D. programs, it’s important for them to gain research experience both to figure out if it’s something they enjoy and to strengthen their resume,” says Associate Professor James Hays, who manages the UROC program.
“Undergraduate research can have larger impact. It can lead to highly visible and influential publications and is sometimes a vehicle for faculty to pursue high risk, high reward projects that don’t yet have funding.”
From undergraduate research opportunities to advanced multi-million dollar interdisciplinary, interagency programs, the GT Computing research community is indeed thriving.
“There’s a difference between vision and reality,” Arkin says.
“Georgia Tech and the College of Computing have strategically invested – and continue to do so – in creating a flourishing research community that builds on our core strengths, embraces diversity of thought, and empowers researchers to discover the next big breakthrough. This is what’s kept me here for more than 30 years.”