GeoAI and Deep Learning Symposium: Big data and GeoAI for Natural Hazards

Type: Paper
Theme:
Sponsor Groups: Cyberinfrastructure Specialty Group
Poster #:
Day: 4/5/2019
Start / End Time: 8:00 AM / 9:40 AM
Room: Washington 6, Marriott, Exhibition Level
Organizers: Qunying Huang, Zhenlong LI, Xinyue Ye
Chairs: Qunying Huang

Call for Submissions

The big data and GeoAI for natural hazards session explores and captures the innovative machine learning and data mining algorithms, techniques and approaches to generate various useful information and products (e.g., hazard maps), which in turn can assist in disaster management during a natural hazard. Topics of particular interest are, but are not restricted to:
1. Mining and extracting actionable information for rapid emergency response and relief coordination
2. Integrating data mining (machines) and crowdsourcing (human) to support decision-making
3. Topic modeling and event detection
4. Physical infrastructure (e.g., roads, bridges and buildings) feature extraction with deep learning
5. Damage assessment using very high-resolution images or/and social media data
6. Coding/classifying text messages during a natural hazard
7. Identifying or/and matching the needs of people in impacted communities
8. Spatiotemporal mining of social media data to gain geographic situational awareness during a disaster
9. Mining, mapping and visualizing public’s behaviors, opinions or sentiments towards a disaster event
10. Innovative approaches to synthesizing and mining multi-sourced social and physical sensing data for disaster management


Description

Recently, we have unfortunately witnessed a series of deadly hurricane events (e.g., Harvey, and Florence) and Northern California wildfires. Such events claim many lives, cause billions of dollars of damage to properties, and severely impact the environment. When a natural hazard occurs, managers and responders need timely and accurate information on damages and resources to make effective response decisions and improve management strategies. This information is referred to as “Situational Awareness” (SA), i.e., an individually as well as socially cognitive state of understanding “the big picture” during critical situations. Fortunately, the popularity and advancement of social and physical sensor networks offer various real-time big data streams for establishing SA. For example, sharing information such as texts, images, and videos through social media platforms enables all citizens to become part of a large sensor network and a homegrown disaster response team. The use of Unmanned Aerial Vehicle (UAVs) images is offering increasing opportunities in disaster-related situations. However, such massive and rapidly changing data streams present new grand challenges to mine actionable data and extract critical validated information for various disaster management activities.

This session explores and captures the innovative machine learning and data mining algorithms, techniques and approaches to generate various useful information and products (e.g., hazard maps), which in turn can assist in disaster management during a natural hazard. Topics of particular interest are, but are not restricted to:
1. Mining and extracting actionable information for rapid emergency response and relief coordination
2. Integrating data mining (machines) and crowdsourcing (human) to support decision-making
3. Topic modeling and event detection
4. Physical infrastructure (e.g., roads, bridges and buildings) feature extraction with deep learning
5. Damage assessment using very high-resolution images or/and social media data
6. Coding/classifying text messages during a natural hazard
7. Identifying or/and matching the needs of people in impacted communities
8. Spatiotemporal mining of social media data to gain geographic situational awareness during a disaster
9. Mining, mapping and visualizing public’s behaviors, opinions or sentiments towards a disaster event
10. Innovative approaches to synthesizing and mining multi-sourced social and physical sensing data for disaster management


Agenda

Type Details Minutes Start Time
Presenter Brady Woods*, University of Maryland - College Park, Kathleen Stewart, University of Maryland - College Park, Region Approximation using Delaunay Filtration and Crowd-Sourced Data During a Natural Disaster 20 8:00 AM
Presenter Yuqin Jiang*, University of South Carolina, Zhenlong Li, University of South Carolina, Analyzing Human Mobility Patterns during Hurricane Matthew Evacuation using Twitter 20 8:20 AM
Presenter Jingchao Yang*, George Mason University, Manzhu Yu, George Mason University, Han Qin, Ankura Consulting Group, Mingyue Lu, Nanjing University of Information Engineering, Chaowei Yang, George Mason University, A Twitter Data Credibility Schema Framework - Hurricane Harvey as a Use Case 20 8:40 AM
Presenter Arif Masrur*, Pennsylvania State University, Human-in-the-loop Machine Learning Framework for Discovering Interesting Association Patterns in Space-Time Big Data 20 9:00 AM
Presenter Qunying Huang*, University of Wisconsin - Madison, Real Time Situational Awareness Information Extraction and Exploration from Social Media 20 9:20 AM

To access contact information login