For further details, please contact:

Dr. Sophie Dale

Project Manager

University of Nottingham

ubicomp at nottingham.ac.uk

Tel: 0115 84 68923

     

Ubiquitous Computing at a Crossroads Workshop

 

Ubiquitous Computing at a Crossroads: Art, Science, Politics and Design 

January 6th and 7th 2009
Huxley Building, Imperial College London

 

Bayesian methods  in Privacy and Trust

Catuscia Palamidessi, INRIA-Futurs and LIX

 

Presentation slides

 

Understanding how unknown information can be inferred from observable data is fundamental in many fields of Security, and in particular in Information Flow, Privacy and Trust. In Information Flow and in Privacy it helps avoiding information leaks that may reveal the secrets. In Trust it is useful to estimate the reliability 
of a user. In an abstract setting, we consider the case in which a random variable X, with unknown distribution, can be tested via repeated independent experiment. The observable results of the experiments can be either the values of X itself (simple case, useful for Trust applications), or the values of another random variable Y which depends  probabilistically on X (case in which X is the input and Y is the output of an protocol, useful for privacy applications). We investigate various approaches to the problem of deriving the distribution of X, and we present a metric framework for comparing the approaches.

 e