CSIS Seminar

Advances in Adversarial Risk Analysis

Abstract

In the first part of my talk I will talk about Adversarial Classification. In multiple domains such as malware detection, automated driving systems, or fraud detection, classification algorithms are susceptible to being attacked by malicious agents willing to perturb the value of instance covariates in search of certain goals. Such problems pertain to the field of adversarial machine learning and have been mainly dealt with, perhaps implicitly, through game-theoretic ideas with strong underlying common knowledge assumptions. These are not realistic in numerous application domains in relation to security. We present an alternative statistical framework that accounts for the lack of knowledge about the attacker’s behavior using adversarial risk analysis concepts. In the second part I will discuss an adversarial risk analysis framework for the software release problem. A major issue in software engineering is the decision of when to release a software product to the market. This problem is complex due to, among other things, the uncertainty surrounding the software quality and its faults, the various costs involved, and the presence of competitors. A general adversarial risk analysis framework is proposed to support a software developer in deciding when to release a product and showcased with an example.

Speaker Bio

Fabrizio Ruggeri (B.Sc. Mathematics Milano, M.Sc. Statistics Carnegie Mellon, Ph.D. Statistics Duke) is Senior Fellow at the Istituto di Matematica Applicata e Tecnologie Informatiche (IMATI) in Milano of CNR (Consiglio Nazionale delle Ricerche) where he had been a researcher from 1988 to May 2023 (as Research Director since 2001). He is member of the Faculty of the Ph.D. programme in Mathematics at the universities of Milano-Bicocca and Pavia. He had various international appointments, including Adjunct Professor at Queensland University of Technology (Brisbane, Australia), International Professor Affiliate at Polytechnic Institute (New York University, USA), Chair of Excellence at Universidad Carlos III and ICMAT-CSIC (Madrid, Spain) and Faculty of the Ph.D. programme in Statistics at the Universidad de Valparaiso (Chile). He is the President-Elect of the International Statistical Institute (ISI) for the 2023-2025 term, after which he will serve as President from 2025 to 2027. He has been President of ENBIS (European Network for Business and Industrial Statistics), ISBA (International Society for Bayesian Analysis) and ISBIS (International Society for Business and Industrial Statistics), and ISI Vice President. He is a Fellow of IMS (Institute of Mathematical Statistics), ASA (American Statistical Association) and ISBA (which also awarded him the first Zellner Medal), and ENBIS Honorary Member. He is Editor-in-Chief of Applied Stochastic Models in Business and Industry and Wiley StatsRef, an online encyclopedia. He is the Director of the Applied Bayesian Statistics summer school organized by CNR-IMATI since 2004 and Chair of the series of workshops on Bayesian Inference in Stochastic Processes. He is a member of the Advisory Scientific Committee of the Institute of Statistics of Academia Sinica, Taiwan. He is author of over 200 articles (including 120 in refereed journals) and author/editor of 6 books. His interests are mostly in Bayesian Statistics and Decision Analysis, especially about robustness, stochastic processes and industrial applications, mainly in reliability. His interests also cover other areas, in particular concentration functions, wavelets and, more recently, healthcare frauds and Adversarial Risk Analysis.