Mathias Humbert
ETH Zurich
67 Papers
304 Citations
Mathias Humbert is an academic researcher from ETH Zurich. The author has contributed to research in topics: Computer science & Inference attack. The author has an hindex of 20, co-authored 59 publications. Previous affiliations of Mathias Humbert include University of Lausanne & armasuisse.
Chat about Author
Papers
ML-Leaks: Model and Data Independent Membership Inference Attacks and Defenses on Machine Learning Models
Ahmed Salem,Yang Zhang,Mathias Humbert,Pascal Berrang,Mario Fritz,Michael Backes +5 more
- 24 Feb 2019
TL;DR: In this article, the authors present the most comprehensive study so far on this emerging and developing threat using eight diverse datasets which show the viability of the proposed attacks across domains, and propose the first effective defense mechanisms against such broader class of membership inference attacks that maintain a high level of utility of the ML model.
1K
FairTest: Discovering Unwarranted Associations in Data-Driven Applications
Florian Tramèr,Vaggelis Atlidakis,Roxana Geambasu,Daniel Hsu,Jean-Pierre Hubaux,Mathias Humbert,Ari Juels,Huang Lin +7 more
- 26 Apr 2017
TL;DR: FairTest as discussed by the authors is a framework for the discovery of unfair, discriminatory, or offensive user treatment in data-driven applications, which is based on the Unwarranted Association (UA) framework.
212
Addressing the concerns of the lacks family: quantification of kin genomic privacy
Mathias Humbert,Erman Ayday,Jean-Pierre Hubaux,Amalio Telenti +3 more
- 04 Nov 2013
TL;DR: This work formalizes the problem and detail an efficient reconstruction attack based on graphical models and belief propagation, and introduces the quantification of health privacy, specifically the measure of how well the predisposition to a disease is concealed from an attacker.
When Machine Unlearning Jeopardizes Privacy
TL;DR: This paper proposes a novel membership inference attack that leverages the different outputs of an ML model's two versions to infer whether a target sample is part of the training set of the original model but out of theTraining set of a corresponding unlearned model.
walk2friends: Inferring Social Links from Mobility Profiles
Michael Backes,Mathias Humbert,Jun Pang,Yang Zhang +3 more
- 30 Oct 2017
TL;DR: Zhang et al. as discussed by the authors proposed a novel social relation inference attack that relies on an advanced feature learning technique to automatically summarize users' mobility features and predict any two individuals' social relation, and it does not require the adversary to have any prior knowledge on existing social relations.
121