El próximo viernes 7 de marzo, a las 10:30, el Dr. Ali Miri, de la Universidad de Otawa, impartirá el seminario titulado Privacy-preserving data mining.

Un resumen de la temática del mismo puede leerse a continuación:

Although, data is very valuable in every organization, it must be processed in order to be useful. Data mining is a collection of techniques which find patterns and associations in raw data, and or classify or cluster the items according to their attributes. Nowadays, related data is normally distributed among two or more parties in different configurations, and mining can be done in an accurate and useful way for all parties involved, on all data collections. However, privacy is often crucial in different scenarios, and organizations involved do not want to disclose their own private information to each other. Therefore, standard algorithms of data mining must be modified, or new protocols have to be designed to preserve the privacy of the parties. In this talk, we will introduce several new Secure Multi-Party Computation (SMC) protocols, and show how they can be used to construct privacy-preserving ID3 decision tree, k-mean clustering algorithm, and neural networks. Time permitting, we will also discuss their computational and communication complexity overheads.