image: Long Island Expressway in New York City shut down due to flash flooding from Post-Tropical Storm Ida's landfall. view more
Credit: Tommy Gao, CC BY-SA 4.0, via Wikimedia Commons
In the event of a natural disaster like Hurricane Ida, artificial intelligence (AI) may be used to assess damage, using imagery reports to identify the severity of flooded areas. Using AI in disaster scene assessment has its limitations, however, and input from the people affected is needed, in order to get a better picture. A new project being led by Dong Wang, associate professor in the School of Information Sciences at the University of Illinois Urbana-Champaign, will explore the power of human intelligence to address the failures of existing AI schemes in disaster damage assessment applications and boost the performance of the system. Wang has received a three-year, $499,786 National Science Foundation (NSF) Human-Centered Computing (HCC) grant for his new project, "DeepCrowd: A Crowd-assisted Deep Learning-based Disaster Scene Assessment System with Active Human-AI Interactions."
"A key limitation of AI-based techniques is the black-box nature of many contemporary models and the consequent lack of interpretability of the results and failures,” he said. “This project investigates the problem of troubleshooting, tuning, and eventually improving the black-box AI algorithms by integrating human intelligence with machine intelligence through active crowd-AI interactions."
According to Wang, the results of the project will lead to more opportunities to fully explore the wisdom from the crowd in various crowd-assisted AI application domains, not only in disaster assessment but also in other areas such as transportation, education, and healthcare. In addition, for the educational component of the NSF grant, this project will give students in STEM and from underrepresented groups a chance to study the interaction between AI and humans.
Wang's research interests lie in the areas of human-centered AI, social sensing, big data analytics, and human cyber-physical systems. His work has been applied in a wide range of real-world applications such as misinformation detection, social network analysis, crowd-based disaster response, intelligent transportation, urban planning, and environment monitoring. He holds a PhD in computer science from the University of Illinois Urbana Champaign.