- Seminario: Fair and Private Learning from Data
- Mer, 19. Settembre 2018, 12:00 h
Fair and Private Learning from Data
Machine learning based technologies are nowadays affecting and extracting information from many aspects of our lives. For this reason machine learning, apart from being able to extract useful information from data, has now to be able to guarantee the privacy of the data and be fair with respect to minority groups.
One way to preserve privacy is to corrupt the learning procedure with noise without destroying the information that we want to extract. Differential Privacy (DP) is one of the most powerful tools in this context and addresses the apparently self-contradictory problem of keeping private the information about an individual observation while learning useful information about a population. From one side it is possible to prove that a learning algorithm which shows DP properties also generalises. From the other side, if an algorithm is not DP, it is possible to state the conditions under which a hold out set can be reused without risk of false discovery through a DP procedure called Thresholdout.
For what concerns the algorithmic fairness in machine learning, the central question is how to enhance supervised learning algorithms with fairness requirements, namely ensuring that sensitive information (e.g. knowledge about ethnic group) does not `unfairly' influence the outcome of a classifier. In this framework it is possible to define new learning paradigms and derive both risk and fairness bounds that support their statistical consistency.
Luca Oneto was born in Rapallo, Italy in 1986. He received his BSc and MSc in Electronic Engineering at the University of Genoa, Italy respectively in 2008 and 2010. In 2014 he received his PhD from the same university in School of Sciences and Technologies for Knowledge and Information Retrieval with the thesis "Learning Based On Empirical Data". He is currently an Assistant Professor in Computer Engineering at University of Genoa with particular interests in Statistical Learning Theory and Data Science.