Big Data All Day
Columbia’s Data Science Institute held its second annual all-day conference devoted to big data on April 6, and it was information overload in the best sense.
Attendees got a chance to learn all about key research underway by Columbia faculty and researchers who are leading the charge in this burgeoning field. Disciplines as diverse as engineering, economics, astrophysics, and history were represented throughout the day on data science challenges as varied as gang violence prevention, measurements of urban pollution, kidney disease prediction, encrypted search, and much more.
Speaking to a packed audience in Roone Arledge Auditorium, Columbia Engineering Dean Mary C. Boyce underscored the Institute’s crucial role since its inception four years ago in pushing the innovation, discussion, and education of data science, across multiple fields. “The Institute has really acted to catalyze data science across the University,” she said, stressing that the initiative has ignited faculty and student collaboration. “We’re also transforming the way we do research and the way we teach across campus,” she added, pointing to new programs in data science, including the MS program, a certificate program for professionals, and several new courses in data science.
In his keynote address, Sidewalk Labs CEO Dan Doctoroff gave a riveting talk about “the coming technological revolution in cities.” Doctoroff, former CEO of Bloomberg L.P. and a former deputy mayor of New York City under Michael Bloomberg, started Sidewalk Labs to bridge the gap between technology and cities. He said harnessing data science—computing power, sensing, the power of social networks—will have a revolutionary impact on people and our urban environment, essentially developing the smartest of all cities that will lead humanity into the future.
Doctoroff noted that we are now in an urban digital network revolution, saying “Technology has not yet integrated into our physical environment. It’s when technology is integrated into the physical environment that we see the greatest changes.”
He believes true connectivity is the future. “When you consider efficiency, adaptability, accountability, personalization, and community, and the impact in combination that those technologies are going to have, I do believe that when we look back 50 years from now … we’re not even going to recognize the city of 2016.”
Data Science Day also provided several lightning talks by Columbia faculty. Engineering Professors Shih-Fu Chang, Julia Hirschberg, Kathleen McKeown, Shipra Agrawal, Hod Lipson, Fred Jiang, and Tal Malkin discussed their research, ranging from projects on mining images, speech, and text for insights and important events to measuring and addressing social and environmental problems in cities.
Attendees surveyed a cross section of the cutting-edge research at Columbia advancing data science, and talked shop with faculty, student innovators, entrepreneurs, and industry affiliates, including Cisco and GE, in a science fair format packed with hands-on demos.
A group of Columbia Engineering students showcased a Robotic Spine Exoskeleton, a collaboration with researchers at Columbia University Medical Center and local hospitals on using sensor data to automatically calibrate three motorized rings for targeted treatment of scoliosis. Electrical Engineering PhD candidate Caroline Yu demonstrated her work using substrates to develop tactile sensing for robots, or “robot skin.”
Other work on display ranged from a smartphone-based structural health monitoring program from civil engineering graduate student Ekin Ozer for crowdsourcing data on structures’ vibrations to innovative security software known as “Heisenbytes” that cleverly deletes code often used to hijack computers just as viruses begin an attack, developed by Computer Science graduate student Adrian Tang and Professor Salvatore Stolfo.
Several projects emerged from the research groups of Electrical Engineering Professors Gil Zussman and Harish Krishnaswamy, investigating how to maximize potential use of scarce spectrum for better wireless communication. One project, FlexICon, is working on full duplex radios capable of canceling out interference to be able to transmit and receive simultaneously, expected to be a key component of 5G technology. Another, AMuSe, is developing a system for wirelessly sending live videos to numerous devices in crowded venues (for example, sport stadiums).
—by Melanie A. Farmer and Jesse Adams