At the Experiential AI for Societal Impact lighting-talks, faculty members share a glimpse into a world enhanced by AI. (Photo by Heratch Ekmekian)
In a single hour, hundreds attending Northeastern University’s Institute for Experiential AI (EAI) launch event got a glimpse into a world enhanced by AI. During a whirlwind of enlightening 10-minute back-to-back talks, called Experiential AI for Societal Impact, chaired by EAI director Jennifer Dy, the institute’s faculty unveiled the powerful potential of these predictive tools to improve society in their respective fields. At this Discover EAI event, the five EAI member presenters, chosen among more than 90 institute members, shared how their algorithms are already improving society in vastly differing spaces, like cybersecurity, climate and sustainability, health care data, developmental disorders, and even humanitarian aid.
AI for Helping the UN Feed Those Most in Need
For the United Nations (UN) World Food Programme, AI has saved millions of dollars and sped up the delivery of food to individuals most in need. That’s right, even the largest humanitarian organization in the world is already using AI, and it was EAI’s very own Ozlem Ergun, College of Engineering distinguished professor, who made it happen.
When Ergun began with the UN’s food assistance program almost a decade ago, all functions of its logistics center were siloed. Every department from those deciding who should receive food to how they received it operated separately.
“This is a very inefficient way of doing things,” says Ergun, adding that the operations of the assistance program are integral to the UN’s bottom line since it provides food and supplies to 80 countries in the world. “To do all these things together you need to have a framework and tools that enable them to make joint decisions.”
Now, a decade later, since Ergun rolled out the organization’s first big data project, AI is identifying, organizing, storing, and sharing crucial data like clockwork across the organization. Her nimble mathematical models identify and propose an optimized supply chain setup so decision-making can work in concert across all departments. And that’s just one of many humanitarian projects by Ergun, who’s also worked on supply chain optimization tools for USAID’s Food for Peace, the European Medicines Agency, and the Centers for Disease Control and Prevention, among others.
AI for Predicting Climate Change
In another riveting lightning discussion, Northeastern professor of civil and environmental engineering Auroop Ganguly, who doubles as director of the Sustainability & Data Sciences Lab, explained how he’s also utilizing AI to bring greater good to the world, though through a very different capacity.
Ganguly is working to ensure infrastructure resilience under extreme weather conditions, often caused by climate change. Leaning on his background in machine learning and nonlinear physics, he discussed the opportunities and challenges of modeling datasets to advance earth systems in cities, despite financial barriers. The takeaway? These solutions, which often involve implementing sensors across municipalities, need to be developed in tandem with incentives and social impacts. Ganguly is advancing our knowledge on behalf of the National Science Foundation, Department of Defense, Department of Homeland Security, NASA, the city of Boston, and even his own Boston-based climate analytics startup, risQ,
AI for Aggregating Medical Text to Find a Cure
As for speaker and associate professor Byron Wallace, he uses AI’s natural language processing to decipher the massive sets of healthcare data in clinical journals and notes in electronic health records. His goal? To create a structure using bioinformatics to advance medicine.
“We’d like to use natural language processing technologies to help healthcare providers make sense of and utilize this data to improve patient care,” says Wallace, who’s also refining models that help interpret scans for radiologists.
In collaboration with the medical community, Wallace has already compiled in real-time more than 800,000 searchable clinical trials that automatically update and detect patterns in the text of journal articles to assess disease risk or the efficacy of drugs, for example. He’s currently fine-tuning this AI tool, called Trialstreamer, now housed at EAI and used out in the field.
AI for Preventing Developmental Disorders
Director of the Augmented Cognition Laboratory, Sarah Ostadabbas, who doubles as assistant professor of electrical and computer engineering, took the Discover EAI podium to discuss her work at the intersection of computer vision and machine learning, with a particular focus on representation learning and visual perception problems. Her algorithms are already helping experts understand, predict, and detect behavior and motor function to innovate and understand developmental disorders. That means using AI for everything from analyzing a patient’s pose for understanding the effects of Parkinson’s drugs, to studying the behavior of bats to help roboticists to design flying drones.
Most recently, Ostadabbas has made progress in elucidating the motor behaviors and function of infants in an effort to prevent a host of developmental disorders, from autism spectrum to cerebral palsy.
“Human pose estimation has received a lot of attention and success lately, but much of it can’t be translated to infants. If you look at adult poses, you can imagine big data from social media, Hollywood images, and sports are not going to work with infant poses,” says Ostadabbas. “Infants have more complex and unique movement and a different body shape with a different muscle-to-bone ratio.”
Her group is the first to publicly release datasets that apply domain adaptation tools to successfully close that data gap between infants and adults.
“The state-of-the-art computer vision doesn’t work on infants moving, but our model does,” says Ostadabbas. “Hopefully it can come full picture and be in every baby monitor out there.”
AI for Stopping Cyber Attacks
Khoury College of Computer Sciences associate professor Alina Oprea rounded out the lightning talks by switching gears to a timely topic of the digital age. Her expertise falls in the area of cyber security, following a growing number of famous attacks in 2011, 2017, and more recently, last year: A computer breach in the oil pipeline system that carries jet fuel and gasoline to the southeastern United States, known as the Colonial Pipeline.
“These attacks equate to financial losses, loss of sensitive and private information, and critical infrastructure threats,” says Oprea. “The question that motivates my research is, can we use AI to solve these security problems?”
Through her project with the Defense Advanced Research Projects Agency, Oprea successfully modeled algorithms to detect threats, both universally to individual organizations, faster and more accurately than existing defenses. These Department of Defense projects are called P-CORE and Portfiler.
“One system has been deployed (and verified by security teams) in two university networks to identify new malware that was not detected by existing tools,” says Oprea. “They even discovered new attacks we didn’t know about.”
Watch more lively multidisciplinary presentations and panel discussions from the April 6 DEAI event, which celebrated the launch of this foundational center for research and applied AI solutions. Contact us now to join forces with EAI in research or business.