|
Professor Kurose talked
about CASA (Collaborative Adaptive Sensing of the Atmosphere), a case-study of
a data-driven sense-and-response system. He elaborated on problems and
solutions in collecting, analyzing and presenting data in large sensor
networks with a specific application to meteorological sensing.
Professor Muthukrishnan
talked about algorithms for analyzing massive streams of data. He provided
insight into data stream models, sublinear space/time algorithms, compressed
sensing and massive distributed systems. He illustrated the algorithms with
several interesting examples in IP network traffic analysis, web traffic
analysis and signal processing.
Professor Raghu talked
about the recent trends in databases with a special emphasis on data mining
and exploratory analysis. He covered a variety of topics such as DBMS support
for complex data analysis (OLAP, warehousing, view materialization), database
support for exploratory data mining, and web data management.
Professor Alex Smola gave a
detailed picture of active topics in machine learning such as Support Vector
Machines (SVM) and other kernel-based learning methods. He provided an
overview of support vector classification and regression, novelty detection,
and quantile regression. He introduced several kernel-based methods and
illustrated them with applications to named entity recognition and ranking of
web pages for search engines.
The topics covered at TECS
Week 2007 introduced the participants to a wide range of issues in
data-intensive computing, some highly theoretical and some highly
application-oriented. Yet, a big picture integrating these varied issues
seemed to emerge at towards the end of the workshop, which participants could
greatly appreciate.
|