SoSySec seminar
Software and Systems Security
Inria - Rennes
December 1st, 2023, 11:00
In-person: Turing-Petri
Remotely with BBB: https://bbb.inria.fr/all-t0p-qjq-9em / Access code: 192737
----------------------------------------------------------------------
Mohamed Maouche (Inria)
----------------------------------------------------------------------
======================================================================
Privacy Challenges in the Era of Deep Learning: Risks and Solutions
======================================================================
The surge of deep-learning systems has developed an imperative to construct large
datasets for their training. However, this growth in data collection also brings forth
significant privacy concerns that the entire data pipeline at risk. These concerns
manifest from the initial data acquisition stage, where the risk of data owner
identification and sensitive information extraction emerge, to the final model
deployment phase, where the model itself can be leveraged for inference attacks,
including membership, reconstruction, and attribute inference.
In this talk, we examine the diverse array of privacy risks inherent in this data pipeline
and present a range of proposed solutions. These solutions are presented through
practical use cases, such as speech anonymization within Automatic Speech
Recognition (ASR) systems and decentralized collaborative filtering for recommender
systems. Additionally, we delve into the distinctions between one-to-one
anonymization techniques and synthetic data generation methods.
----------------------------------------------------------------------
To follow the presentation remotely, please connect to the following
URL with a modern web browser:
- URL: https://bbb.inria.fr/all-t0p-qjq-9em
- Access code: 192737
----------------------------------------------------------------------
Seminar taking place in person with mandatory registration at least
*72h beforehand* for *all* in-person participants by email to
Nadia Derouault<nadia [*] derouault
inria [*] fr>. Participants
non-affiliated with Inria or IRISA will be asked to present an ID at
the reception desk of the IRISA building.
To receive the SoSySec announcements, please subscribe to the SoSySec
mailing list:
https://sympa.inria.fr/sympa/subscribe/sosysec
All past and future SoSySec talks are listed at
https://seminaires-dga.inria.fr/en/seances-a-venir/
----------------------------------------------------------------------
Séminaire en présentiel ouvert à tous mais avec inscription
obligatoire au moins 72h à l'avance pour *tous* les participants
en présentiel auprès de Nadia Derouault<nadia [*] derouaultinria [*] fr>.
Les participants externes devront se présenter à l'accueil avec une
pièce d'identité.
Vous pouvez vous abonner à nos annonces de séminaires :
https://sympa.inria.fr/sympa/subscribe/sosysec
et consulter la liste des exposés passés et à venir :
https://seminaires-dga.inria.fr/seances-a-venir/