GeoAI and Deep Learning Symposium: Geo-Text Data and Location-based Social Media

Type: Paper
Theme:
Sponsor Groups: Geographic Information Science and Systems Specialty Group, Spatial Analysis and Modeling Specialty Group, Cyberinfrastructure Specialty Group
Poster #:
Day: 4/4/2019
Start / End Time: 1:10 PM / 2:50 PM
Room: Capitol Room, Omni, East
Organizers: Yingjie Hu, Roger Wang, Gengchen Mai
Chairs: Christa Brelsford

Call for Submissions


Example topics (include but not limited to):
- Spatial, temporal, and semantic analytics on geotagged social media data
- Deep learning and related techniques for extracting and disambiguating place names from texts
- Deep learning and related techniques for geo-text data classification
- Sentiment analysis and geospatial clustering based on geo-text data
- Topic modeling and geospatial analysis with geo-text data
- Gazetteer construction based on geo-text data analysis
- Statistical models for geo-text data
- Applications of spatial, temporal, and textual data analysis
- …


To present your research in our session, please submit your abstract through AAG website (http://www.aag.org/cs/annualmeeting/register) and send your PIN to Yingjie Hu (yhu42@buffalo.edu), Christa Brelsford (brelsfordcm@ornl.gov), or Roger Wang (wangruoqian@gmail.com) by October 25, 2018.


Organizers:
Yingjie Hu, University at Buffalo
Christa Brelsford, Oak Ridge National Lab
Roger Wang, Rutgers University


Description

Datasets containing natural language text that is linked to geographic location information are changing our lives. Examples include location-based social media, geotagged Wikipedia pages, geotagged housing advertisements, news articles, historical archives, business documents, and more. In traditional geographic data, e.g. temperature measurements or digital elevation models, numeric values are assigned to locations, which are relatively easy to process and extract information. In comparison, processing geo-text data requires more careful analysis, for two reasons: first, the text data are often unstructured, and inferring consistent information from the textual data is challenging. Second, the relationship between the text and location can be implicit rather than explicit. This means that careful attention needs to be paid to uncertainty, so that our understanding isn’t biased by data misinterpretation. The advance of GeoAI and deep learning provides new opportunities to address this issue. They enable the extraction of a variety of information from unstructured texts including named entities, topics, emotions and many others. Combined with geographic locations, the extracted information enables numerous spatial and spatiotemporal applications. For example, a combination of geo-text data with deep learning and other AI techniques have been used in addressing problems in disaster response, land use mapping, public health, population distribution, and others. Meanwhile, there is also a demand for more theories in order to explain and understand the developed computational models and the obtained analysis results. This special session aims to bring together researchers with interests in geo-text data and to provide an ideal venue to present research results, exchange ideas, and envision the future of this field.


Agenda

Type Details Minutes Start Time
Presenter Pengyuan Liu*, University of Leicester, Stefano De Sabbata, University of Leicester, Learning Digital Geographies through a Correlation-based Autoencoder 20 1:10 PM
Presenter Roger Wang*, Rutgers University, Social Media Diffusion Pattern Recognition for Flood Hazards using Dimensionality Reduction 20 1:30 PM
Presenter Yingjie Hu*, University At Buffalo, EUPEG: Towards an Extensible and Unified Platform for Evaluating Geoparsers 20 1:50 PM
Presenter Gengchen Mai*, University of California - Santa Barbara, CA, Bo Yan, Univerity of California - Santa Barbara, CA, Krzysztof Janowicz, University of California - Santa Barbara, CA, Rui Zhu, University of California - Santa Barbara, CA, Relaxing Unanswerable Geographic Questions Using A Spatially Explicit Knowledge Graph Embedding Model 20 2:10 PM
Presenter Xi Liu*, Pennsylvania State University, Nuts and Bolts of GeoAI research for understanding our geo-social systems 20 2:30 PM

To access contact information login