AI for Extreme Event Reslience – Ibrahim Demir

Knowledge Discovery, Integration, and Communication for Extreme Weather and Flood Resilience Using Artificial Intelligence: Flood AI Alpha

IS-GEO Lecture Series

Flood AI Alpha



Dr. Ibrahim Demir, Assistant Professor, University of Iowa

Presentation Slides: Flood_AI Slides_byIbrahimDemirApril2017

Audio: (link in prep) |Video link

Learn more about Dr. Demir’s research at:

Presented at: IS-GEO Monthly Telecon | April 17, 2017

Abstract: Nobody is immune from extreme events or natural hazards that can lead to large-scale consequences for the nation and public. One of the solutions to reduce the impacts of extreme events is to invest in improving resilience with the ability to better prepare, plan, recover, and adapt to disasters. The National Research Council (NRC) report discusses the topic of how to increase resilience to extreme events through a vision of a resilient nation in the year 2030. The report highlights the importance of data, information, gaps and knowledge challenges that need to be addressed, and suggests every individual to access the risk and vulnerability information to make their communities more resilient. This abstract presents our project on developing a resilience framework for flooding to improve societal preparedness with objectives; (a) develop a generalized ontology for extreme events with a primary focus on flooding; (b) develop a knowledge engine with voice recognition, artificial intelligence, natural language processing, and inference engine. The knowledge engine will utilize the flood ontology and concepts to connect user input to relevant knowledge discovery outputs on flooding; (c) develop a data acquisition and processing framework from existing environmental observations, forecast models, and social networks. The system will utilize the framework, capabilities and user base of the Iowa Flood Information System (IFIS) to populate and test the system; (d) develop a communication framework to support user interaction and delivery of information to users. The interaction and delivery channels will include voice and text input via web-based system (e.g. IFIS), agent-based bots (e.g. Microsoft Skype, Facebook Messenger), smartphone and augmented reality applications (e.g. smart assistant), and automated web workflows (e.g. IFTTT, CloudWork) to open the knowledge discovery for flooding to thousands of community extensible web workflows.