Novel Methods in Data Summarization, Anonymization, and Indexing

Computer Science and Engineering
NYU Community Event


Panagiotis Karras
National University of Singapore


This talk will present highlights of my research in three areas of data management and mining. The first area concerns data summarization under space or accuracy constraints. I will mention instances of novel algorithms and structures for synopsis construction I have introduced, which outperform previous state-of-the-art in both accuracy and efficiency, while going beyond conventional assumptions in the area to show how some problems are not as hard as previously thought. The second area concerns the transformation of relational and transaction (set-valued) data in order to satisfy a privacy constraint. I will shortly outline two such instances, namely an algorithm that satisfies the l-diversity model, and a novel model for transaction data publishing with algorithms therefor. Last, I will discuss the problem of indexing and answering complex queries on semi-structured Semantic Web data,and the Hexastore, a sextuple index structure I introduced for that purpose. The focus will be on the core ideas on each topic, the pedagogic insights one can find in them. The end of the talk will outline the directions of my future research plans.

About the Speaker

Panagiotis Karras is an LKY Postdoctoral Fellow at the National University of Singapore. He received an MEng in Electrical and Computer Engineering from the National Technical University of Athens, a PhD in Computer Science from the University of Hong Kong, and the 2008 Young Scientist Award in Physical/Mathematical Science from the Hong Kong Institution of Science.

In the past he has also worked and studied at the University of Zurich, the Technical University of Denmark, the Institute of Language and Speech Processing in Athens, Schlumberger Information Solutions in Oslo, the University of Karlsruhe, Germany, and the University of Patras, Greece.