Output details
11 - Computer Science and Informatics
University of Birmingham
A Statistical Test for Information Leaks Using Continuous Mutual Information
<10>This work describes a framework for calculating side channel leakage using continuous mutual information. The use of the discrete version is well understood, but when this measure is applied to systems with continuous observations (such as time measurements) much of the useful data can be lost. Our contribution was to calculate the distribution that estimates of information leakage based on discrete and continuous mutual information would come from and show that using continuous measures was significantly more accurate. This paper led directly to the ESPRC grant "New Techniques for finding information leaks" EP/J009075/, which was funded in 2012.