Seminars & Colloquia
Oak Ridge National Laboratory
"Analytics for Big Spatial and Spatiotemporal Data: Algorithms, Applications, and Outlook"
Monday November 11, 2013 09:30 AM
Location: 3211, EBII NCSU Centennial Campus
(Visitor parking instructions)
Data collection has exploded in recent years owing to advances in sensor technology, the proliferation of location sensing devices such as cell phones, increased collection of user activities on the web, user generated content at social media portals, and increased resolution (space and time) in scientific simulations (climate, transportation). One interesting aspect that is common to many of these diverse datasets, apart from volume, variety, velocity, and veracity, is that they are embedded in space and time. Novel and scalable algorithms are required in order to extract useful geo-information from these ever-growing multi-sensor and multi-resolution datasets. In this talk, I will focus on the challenges posed by high-resolution (both spatial and temporal) and multi-sensor remote sensing imagery. In particular, I will present recent algorithmic advances in pixel-based classification (semi-supervised learning), object-based classification (multiple-instance learning), image-based visual categorization (semantic classification), and multi-temporal change detection. Scaling some of these methods for large-scale spatiotemporal datasets for regional and global applications in the area of settlement mapping, biomass monitoring, and national security will also be presented. I will conclude by presenting emerging applications and research themes around big spatial and spatiotemporal data.
Dr. Raju Vatsavai is the Lead Data Scientist in the Computational Sciences and Engineering Division at the Oak Ridge National Laboratory, where his primary research for the past seven years has been focused on large-scale spatial and spatiotemporal data management and data mining. He has published over 70 peer-reviewed articles, co-edited two books and a special issue of Intelligent Data Analysis Journal (Volume 13(3), 2009). He is serving on the editorial board (action editor) of Springer’s “GeoInformatica: An International Journal on Advances of Computer Science for Geographic Information Systems.” He co-authored the “Geographic Data Mining and Knowledge Discovery” research priority for the University Consortium on Geographic Information Science (UCGIS), served on the program committees of several prominent international conferences, including ACM SIGKDD Conference on Knowledge Discovery and Data Mining, SIAM Data Mining, and ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. He founded/co-founded and co-chaired several workshops, including SensorKDD (with ACM-SIGKDD: 2007-2011), SSTDM (with IEEE ICDM: 2008-2013), KDCloud (with IEEE ICDM: 2010-2013), PDAC (with ACM/IEEE Supercomputing: 2010-2013), HPDGIS (with ACM SIGSPATIAL GIS: 2010-2011), and LDMTA (with ACM-SIGKDD 2011), and he co-organized invited sessions on data mining and machine learning for remote sensing at the leading IEEE IGARSS (2009-2011) conference. He regularly serves on NSF and DOE review panels.
He has also contributed extensively to several highly successful software systems in various capacities (UMN-MapServer - a world leading open source WebGIS, *Miner - a spatiotemporal data mining workbench, EASI/PACE Parallel Fly!, Parallel SAR, and the first parallel softcopy photogrammetry system for IRS-1C/1D satellites). Dr. Vatsavai is the recipient of several awards, including the “Bravo! The Best of IBM!,” the “Best Director’s R&D award for LDRD category at ORNL (2011),” and the Department of Energy’s “outstanding mentor (2013).” His research interests include large-scale spatiotemporal data management and data mining, spatial computing, geospatial intelligence, remote sensing image understanding, and high-performance computing.
Host: Matt Stallmann, Computer Science, NCSU
Back to Seminar Listings
Back to Colloquia Home Page