Machine learning models are known to memorize the unique properties of individual data points in a training set. This memorization capability can be exploited by several types attacks infer information about data, most notably, membership inference attacks. In this paper, we propose an approach based on leakage for guaranteeing privacy. Specifically, use conditional form notion maximal quantify...

Abstract We study a dynamic information design problem in finite-horizon setting consisting of two strategic and long-term optimizing agents, namely principal (he) detector (she). The observes the evolution Markov chain that has states, one “good” “bad” absorbing state, to decide how sequentially disclose detector. detector’s only consists messages she receives from principal. objective is dete...

Marcello A Tonelli, MD, SM, FRCPC (Work Group Co-Chair), received an MD from the University of Western Ontario, specialist certification in nephrology (FRCPC) at Dalhousie University, and an SM in epidemiology from Harvard University. He is a nephrologist and Associate Professor at the University of Alberta. He serves as Associate Editor of American Journal of Kidney Diseases, the Cochrane Rena...

John J V McMurray, MD, FRCP, FESC (Work Group CoChair), is Professor of Medical Cardiology at BHF Glasgow Cardiovascular Research Centre and Head of Section of Academic Cardiology at University of Glasgow. Dr McMurray received his medical degree from University of Manchester and completed additional clinical training in Edinburgh, Dundee and Glasgow. He conducts clinical research in a wide span...

Side-channel attacks utilize information leakage in the implementation of an otherwise secure cryptographic algorithm to extract secret information. For example, adversaries can extract the secret key used in a cryptographic algorithm by observing cache-timing data. Threshold cryptography enables the division of private keys into shares, distributed among several nodes; the knowledge of a subse...

We consider a price discrimination problem in which a seller has a single object for sale to a potential buyer. At the time of contracting, the buyer’s private type is his incomplete private information about his value, and the seller can disclose additional private information to the buyer. We study the question of whether discriminatory information disclosure can be profitable to the seller u...

We study the effect of disclosure requirements in environments where experts publicly acquire private information before engaging in a persuasion game with a decision-maker. In contrast to settings where private information is exogenous, we show that disclosure requirements do not change the set of pure-strategy equilibrium outcomes regardless of the players’ preferences.

John A Kellum, MD, FCCM, FACP (Work Group Co-Chair), is Professor of Critical Care Medicine, Medicine, Bioengineering and Clinical and Translational Science, and Vice Chair for Research within the Department of Critical Care Medicine at the University of Pittsburgh. In addition, he is the Director of the Program on Bioengineering and Organ Support for the CRISMA (Clinical Research Investigation...

We address the problem of controlling information leakage in a concurrent declarative programming setting. Our aim is to define formal tools in order to distinguish between authorized, or declared, information flows such as password testing (e.g., ATM, login processes, etc.) and non-authorized ones. We propose to define security policies as rewriting systems. Such policies define how the privac...