To be held in conjunction with

Home

SensorKDD-
2008 Website
Launched

Feb. 16 , 2008
Knowledge Discovery from Sensor Data to be published in 2008

Oustanding presented papers are being submitted for publication in a top Journal

Invited Speakers

In addition to the oral presentation of accepted papers, there will be four invited speakers:

Invited Speaker Bio-Sketches and Abstracts:

Dr. Pedro Domingos is an Associate Professor of Computer Science and Engineering at the University of Washington, Seattle. He received M.S. and Ph.D. in Information and Computer Science from the University of California at Irvine in 1994 and 1997 respectively. He was an assistant professor at IST for two years, before joining the faculty of the University of Washington in 1999. He is author or co-author of over 100 technical publications in machine learning, data mining, and other areas. He is a member of the editorial board of the Machine Learning Journal and the advisory board of JAIR, and a co-founder of the International Machine Learning Society. His research interests include machine learning and data mining.

A General Framework for Mining Massive Data Streams
(Joint work with Geoff Hulten)

Abstract:

In many domains, data now arrives faster than we are able to mine it.  To avoid wasting this data, we must switch from the traditional "one-shot" data mining approach to systems that are able to mine continuous, high-volume, open-ended data streams as they arrive.

In this talk I will identify some desiderata for such systems, and outline our framework for realizing them. A key property of our approach is that it minimizes the time required to build a model on a stream, while guaranteeing (as long as the data is i.i.d.) that the model learned is effectively indistinguishable from the one that would be obtained using infinite data. Using this framework, we have successfully adapted several learning algorithms to massive data streams, including decision tree induction, Bayesian network learning, k-means clustering, and the EM algorithm for mixtures of Gaussians.

These algorithms are able to process on the order of billions of examples per day using off-the-shelf hardware. Building on this, we have developed VFML, a library of software primitives for scaling arbitrary learning algorithms to massive data streams with minimal effort.

TOP OF PAGE

Dr. Joydeep Ghosh is the Schlumberger Centennial Chair Professor of Electrical and Computer Engineering at the University of Texas, Austin. He received his Ph.D. from the University of Southern California in 1988 and his B. Tech from IIT Kanpur in 1983. He has published more than 200 refereed papers and 30 book chapters, and co-edited 18 books. His research has been supported by the NSF, Yahoo!, Google, ONR, ARO, AFOSR, Intel, IBM, Motorola, TRW, Schlumberger and Dell, among others. He is a Fellow of the IEEE and the founder-director of Intelligent Data Exploration and Analysis Lab (IDEAL). His research interests include intelligent data analysis, data mining and web mining, and their applications to a wide variety of complex engineering and AI problems.

A Probabilistic Framework for Mining Distributed Sensory Data under Data Sharing Constraints

Abstract:

Sensory data is often gathered simultaneously from geographically disparate sources.  Such situations also often impose constraints stemming from data ownership, or computational/memory/power limitations that prevent all the data from being gathered at a central location before standard data mining tools can be applied. Moreover, all data attributes may not be available at each data site.  I will describe a general probabilistic framework that efficiently allows (semi-) supervised learning in such situations without being substantially affected by the domain constraints. Implications for design and analysis of future large-scale distributed sensor data will be discussed.

TOP OF PAGE

Dr. Hillol Kargupta is an Associate Professor in the Department of Computer Science and Electrical Engineering at the University of Maryland Baltimore County. He received his Ph.D. in Computer Science from University of Illinois at Urbana-Champaign in 1996. He has published more than seventy five peer-reviewed articles in journals, conferences, and books. He is an associate editor of the IEEE Transactions on Knowledge and Data Engineering, IEEE Transactions on Systems, Man, Cybernetics, Part B, and Statistical Analysis and Data Mining Journal, among others. Dr. Kargupta is also a co-founder of AGNIK LLC, a ubiquitous data intelligence company. His research interests include mobile and distributed data mining and computation in gene expression.

Data Mining in Vehicular Sensor Networks: Technical and Marketing Challenges

Abstract:

Data intensive sensor networks are starting to emerge in academic literature and commercial applications. Data mining in such sensor networks offer challenges for researchers and practitioners on several grounds – algorithmic, systems, and marketing. Solutions that work in practice often pay close attention to the needs from each of these domains. This talk will offer a perspective of some of these challenges and their solutions using a recently introduced commercial system for data stream mining in vehicular sensor networks. The talk will discuss the application, the system architecture, and some of the key technical challenges for monitoring vehicle-sensor-data-streams. It will connect those with challenges in marketing a new technology. Rest of the talk will discuss some of the solutions, demonstrate the system, and summarize the broader lessons.

TOP OF PAGE

Dr. Brian A. Worley is the Director of the Computational Sciences and Engineering Division (CSED) at the Oak Ridge National Laboratory (ORNL) and has more than 29 years experience as a research staff member at the Oak Ridge National Laboratory. He holds an adjunct faculty appointment at the Georgia Institute of Technology. He received his Ph.D. and M.S. degrees in Nuclear Engineering from the Massachusetts Institute of Technology in 1977 and 1975 respectively, and a B.S. degree from the University of Tennessee in 1973. Dr. Worley began his research at ORNL in 1977 as a reactor physicist in the Neutron Physics Division. From 1990 to 1999, he served as Group Leader for the Reactor Physics Group and led the neutronics work for the Advanced Neutron Source and DOE's New Production Reactor Program. From 1999 through 2001, he served as Section Head for Computational Sciences in the Computer Sciences and Mathematics Division. Dr. Worley has served in his current position as division director for Computational Sciences and Engineering Division since its inception in October 2001. Over the past twelve years, Dr. Worley has also led numerous defense-related projects for the Department of Defense in a wide range of modeling and simulation disciplines. His research interests are now focused on the general area of knowledge discovery from dynamic, disparate data.

New Frontiers and Opportunities in Sensor-driven Knowledge Discovery

Abstract: 

A viewpoint on knowledge discovery trends is presented as applied to the problem space of massive volumes of dynamic, distributed and heterogeneous data obtained from sensors in physical and cyber space.  Emerging national and societal requirements in national security and consequence management are discussed in this context.  New challenges in areas like automated hypothesis generation and real-time knowledge discovery are emphasized.  An overview of the research agenda and accomplishments of the SensorNet program is presented with a view to motivating emerging and new research directions.

TOP OF PAGE

Detailed
Description