IN004
AI/ML Assurance: Applications in Geospatial Sciences
Session ID#: 161001
Session Description:
With the increase in availability of data and emergence of novel AI/ML methodologies, the need to study the trustworthiness and assurance of data/model quality has increased. Questions on data redundancy, quality and reliability are becoming more important as the availability of data from multiple sources has grown. And with significant implications on the human and global level, there has been recent scrutiny on the privacy impact of most research.
Interest in data/model assurance has rapidly increased across multiple application domains of AI/ML (e.g., medicine and self-driving cars). Many of the approaches and techniques for building assurances will be necessary within earth sciences, especially those impacting decision/policy making or with direct impact on humans lives, such as machine learning and AI applications utilized to build climate models, study climate change, analyze satellite imagery, identify forest covers, and model declining ice sheets heights.
The session invites papers in machine learning and artificial intelligence with applications in geo-spatial sciences, earth sciences, climate modeling, remote sensing or other related domains that discuss (i) model robustness, (ii) model reproduciblity, (iii) model trustworthiness and explainability, (iv) data quality and integrity and (v) privacy. We welcome discussions on open challenges and needs for the community, as well as new methods that address these needs and challenges.
Index Terms:
0550 Model verification and validation [COMPUTATIONAL GEOPHYSICS]
1942 Machine learning [INFORMATICS]
1952 Modeling [INFORMATICS]
3275 Uncertainty quantification [MATHEMATICAL GEOPHYSICS]
Primary Convener: Sreelekha Guggilam, Oak Ridge National Lab, Oak Ridge, United States
Conveners: Vandy Tombs, Oak Ridge National Lab, Oak Ridge, United States, Dan Lu, Oak Ridge National Laboratory, Oak Ridge, TN, United States and Abani Patra, Tufts University, Data Intensive Sciences Center, Medford, MA, United States
Student/Early Career Convener: Sreelekha Guggilam, Oak Ridge National Lab, Oak Ridge, United States
Merging Multiple Sensing Platforms and Deep Learning empowers Individual Tree Mapping and Tree Species Detection at a City-scale (1052512)
Ryoungseob Kwon1, Youngryel Ryu2, Tackang Yang1, Zilong Zhong3 and Jungho Im4, (1)Seoul National University, Department of Landscape Architecture and Rural Systems Engineering, Seoul, Korea, Republic of (South), (2)Seoul National University, Department of Landscape Architecture and Rural Systems Engineering, Seoul, South Korea, (3)Seoul National University, Research Institute of Agriculture and Life Sciences, Seoul, South Korea, (4)Ulsan National Institute of Science and Technology, Urban and Environmental Engineering, Ulsan, South Korea
The Robustness Evaluation Of The Semantic Segmentation Model In High-resolution Satellite Images (1062201)
Tatsuyuki Sekine, Amanogi, Tokyo, Japan, Kenichi Sasaki, University of Colorado Boulder, Boulder, CO, United States, Britta Fukuoka, Amanogi Corp., Ota, Japan and Yu Kudo, Amanogi Corp., Tokyo, Japan
Adopting Grad-CAM for Deep Learning-Based Rock Classification Model (1065042)
Wanhyuk Seo, Graduate Student, Yonsei University, School of Civil and Environmental Engineering, Seoul, South Korea and Tae Sup Yun, Professor, Yonsei University, School of Civil and Environmental Engineering, Seoul, South Korea
SAM guided Half Siamese network model for endmember variability based spectral unmixing of geological materials (1088403)
Maitreya Mohan Sahoo1, R Kalimuthu1, Arun P.V.2, Shibu K. Mathew3 and Alok Porwal1, (1)Indian Institute of Technology Bombay, Mumbai, India, (2)Indian Institute of Information Technology Sri City, Chittoor, India, (3)Udaipur Solar Observatory, Udaipur, India
Leveraging Remote Sensing and Deep Learning to Estimate Suspended Sediment Concentration (1092168)
Arash Modaresi Rad, Boise State University, Computing, Boise, United States, Hannah Spero, University of Notre Dame, Notre Dame, United States, Kendra Fallon, Boise State University, Department of Geosciences, Boise, ID, United States, Jason Kreitler, USGS Western Geographic Science Center, Boise, United States and Mojtaba Sadegh, Boise State University, Department of Civil Engineering, Boise, ID, United States
Pattern Extraction Through Deep Learning: Road as A Feature Set (1100900)
Devendra Patil1, Madhumita Dey2 and Bharath Haridas Aithal2, (1)Indian Institute of Technology Kharagpur, Ranbir and Chitra Gupta School of Infrastructure Design and Managment, Kharagpur, India, (2)Indian Institute of Technology Kharagpur, Ranbir and Chitra Gupta School of Infrastructure Design and Management, Kharagpur, India
Quantifying the Reliability and Robustness of Random Forest and Boosted Decision Tree Models with Large Training Data Sets (1117122)
Caleb Spradlin1, Mark Carroll2, Zachary Williams2, Roger L. Gill2, Mariana Blanco-Rojas1, Savannah Strong2, Jian Li2, Jordan Caraballo-Vega2 and Glenn Tamkin2, (1)NASA Goddard Space Flight Center, Data Science Group, Greenbelt, MD, United States, (2)NASA Goddard Space Flight Center, Data Science Group, Greenbelt, United States
A Meta-Learning Framework for Characterizing and Accessing Training Data for GLOBE Observer Program Land Cover Protocols (1123989)
Xiao Huang, University of Arkansas, Department of Geosciences, Fayetteville, United States, Shawna McBride, University of Wyoming, Laramie, United States, Di Yang, University of Wyoming, Wyoming Geographic Information Center, Laramie, United States, Jessica J Mitchell, University of Montana, Spatial Analysis Lab, Missoula, MT, United States, Peder Nelson, Oregon State University, Corvallis, United States and Russanne Low, Institute for Global Environmental Strategies, Arlington, United States
Uncertainty quantification in the retrieval of cloud properties for Sentinel-4 and TROPOMI / Sentinel-5 Precursor (S5P) using deep neural networks (1143043)
Fabian Romahn1, Diego G Loyola2, Adrian Doicu3, Víctor Molina García2, Ana del Aguila2, Ronny Lutz2 and Athina Argyrouli4, (1)German Aerospace Center DLR Oberpfaffenhofen, Remote Sensing Technology Institute (IMF), Oberpfaffenhofen, Germany, (2)German Aerospace Center (DLR), Remote Sensing Technology Institute (IMF), Oberpfaffenhofen, Germany, (3)German Aerospace Center DLR Oberpfaffenhofen, Oberpfaffenhofen, Germany, (4)Technical University of Munich (TUM), Department of Aerospace and Geodesy, Chair of Remote Sensing Technology, Munich, Germany
A Deep Learning Approach to Distributed Rainfall-Runoff Modeling with Spatiotemporal Interpretation. (1143108)
Qi Deng, Tulane University of Louisiana, ByWater Institute, New Orleans, LA, United States, Ting Liu, Arizona State University, School of Sustainable Engineering and the Built Environment, Ira A. Fulton Schools of Engineering, Tempe, AZ, United States, Yuhang Wei, Arizona State University, Future H2O, Julie Ann Wrigley Global Futures Laboratory, Tempe, United States and John L Sabo, Tulane University of Louisiana, ByWater Institute, New Orleans, United States
Filling Data Gap in MODIS Daily Aerosol Optical Depth Products by Developing an Ensemble Machine Learning Algorithm (1152623)
Alok Kumar1, Varun Katoch1, Fahad Imam1, Sushovan Ghosh2, Kuldeep Dixit2 and Sagnik Dey1, (1)IIT Delhi, New Delhi, India, (2)Indian Institute of Technology Delhi, New Delhi, India
Towards trustworthy machine learning workflow for mapping gravity anomaly: Preliminary exploration of XAI (1162252)
Lexie Yang1, Jordan Bowman1, Jacob Arndt1, Dalton Lunga2, Jason Wohlgemuth2, Dawn King3 and Clifton F Minter4, (1)Oak Ridge National Laboratory, Geographic Data Science, Oak Ridge, TN, United States, (2)Oak Ridge National Laboratory, Geographic Data Science, Oak Ridge, United States, (3)National Geospatial-Intelligence Agency, Office of Geomatics, St. Louis, United States, (4)National Geospatial Intelligence Agency, Springfield, VA, United States
Mitigation of Spatial Nonstationarity with Vision Transformers (1163241)
Lei Liu1, Javier E. Santos2, Masa Prodanovic3 and Michael Pyrcz1, (1)University of Texas at Austin, Austin, TX, United States, (2)Los Alamos National Laboratory, Los Alamos, United States, (3)The University of Texas at Austin, Hildebrand Department of Petroleum and Geosystems Engineering, Austin, TX, United States
Mixed Land Use Detection via Multi-modal Learning (1168836)
Meiliu Wu, University of Wisconsin Madison, Geography, Madison, WI, United States and Qunying Huang, University of Wisconsin, Madison, Department of Geography, Madison, United States
Application of Machine Learning for PM2.5 Estimation Using the Low-cost Sensors (1177779)
Kuldeep Dixit1, Meng Qi1, Tianjun Lu2 and Steve Hankey1, (1)Virginia Polytechnic Institute and State University, School of Public and International Affairs, Blacksburg, United States, (2)California State University Dominguez Hills, Department of Earth Science and Geography, Carson, United States
