Dr. Slobodan Vucetic, Temple University

TITLE: Memory-Constrained Data Mining

ABSTRACT: Data mining concentrates on the development of computationally
efficient and accurate tools for knowledge discovery from very large data
sets. Many successful data mining algorithms scale linearly or even better
with data size and can learn with high accuracy from millions of examples
on a regular PC. Interestingly, linear or sub-linear scaling is often not
enough in a today's world of ever increasing data collections in sciences,
medicine, engineering, and business at one end and increasing use of
microcontrollers with highly limited memory and processing power at
another end. A critical question in this new world is how to optimize
learning from data that exceeds the available computational resources by
orders of magnitude.
   This talk will demonstrate that highly accurate learning could be
achieved using seemingly inferior computational devices. To accomplish
this, a computer is treated as a reservoir that sequentially observes a
large data stream and, at any given moment, maintains a data summary
and a prediction model that describes the data. After each observed stream
example, the reservoir content is updated such that the accuracy is
increased while the memory constraint is kept satisfied. The proposed
approach prefers the examples on which the current predictor is the most
uncertain. In addition, a data summary is maintained instead of the
individual examples.
  In this talk I will present an online support vector machine algorithm
that requires constant memory and achieves sublinear runtime. The
proposed algorithm was tested on a number of large data sets (some of
them containing more than 1 million examples) from various domains,
where it was assumed that the memory can maintain information about only
100 examples. In most cases, the achieved accuracy was near the upper
bound achievable using unlimited resources. On the other hand, the
accuracy was substantially larger than when randomly selected 100
support for the hypothesis that innovative data mining methods can
enable high-quality learning from large data sets using
resource-constrained computing devices.

BIO: Slobodan Vucetic received B.S. and M.S. degrees from the University of
Novi Sad, Serbia, in 1994 and 1997, and Ph.D. degree in electrical
engineering from the Washington State University in 2001. He is an
assistant professor at the Department of Computer and Information Sciences
at Temple University. His research interests include bioinformatics,
knowledge discovery from sequence, time and spatial data, scientific data
mining, text and web mining, machine learning, and data compression.