Your browser is unsupported

We recommend using the latest version of IE11, Edge, Chrome, Firefox or Safari.

Mar 11 2022

Alpha-information theory, information leakage and side-channel attacks


March 11, 2022

11:00 AM - 12:15 PM


online, contact department for login

Alpha-information theory, information leakage and side-channel attacks

Presenter: Olivier Rioul, Telecom Paris

Abstract: In this presentation, the basic "operational" definitions of information-theoretic quantities such as entropy, relative entropy, mutual information, and Fisher information to solve specific problems such as compression data, data transmission, and estimation from observed data will be reviewed. The distinction between asymptotic achievability results and converse theorems that rely on mathematical inequalities such as data processing and Fano inequalities will be empasized. Other types of problems may require other types of information, other inequalities or other criteria.

Measuring information leakage in cryptographic systems under side-channel attacks will be discussed. In this field, the notions of maximum a posteriori (MAP) as well as guessing entropy are important. To obtain converse theorems, i.e., bounds on the success of any type of attack, “alpha-information theory" is considered, which is based on the notions of Rényi's alpha-entropy and alpha-divergence, Arimoto's conditional alpha-entropy, Sibson's alpha-information, and a recently-defined conditional alpha-information. Alpha-information theory generalizes Shannon's classical theory corresponding to alpha=1, makes the link with Hartley's information for alpha=0, with guessing entropy for alpha=1/2, and with MAP for alpha=+infinity. What matters here is the preservation of essential computing and Fano inequalities for any alpha. These inequalities make it possible to bound the success of any possible attack of a cryptographic system from disclosed measurements (by a side-channel) compared to an a priori belief (without access to the measurements).

Speaker bio: Olivier Rioul is full professor at the Department of Communication and Electronics at Télécom Paris, Institut Polytechnique de Paris, France. He graduated from École Polytechnique, Paris, France in 1987, and from École Nationale Supérieure des Télécommunications, Paris, France in 1989. He obtained his PhD degree from École Nationale Supérieure des Télécommunications, Paris, France in 1993. His research interests are in applied mathematics and include various, sometimes unconventional, applications of information theory such as inequalities in statistics, hardware security, and experimental psychology. He has been teaching information theory at various universities for twenty years and has published a textbook which has become a classical French reference in the field.

Faculty Host: Natasha Devroye (

This seminar will not be recorded.


ECE Student Affairs

Date posted

Mar 10, 2022

Date updated

Mar 10, 2022