Adversarial Threats on Real Life Learning Systems
Workshop on Machine Learning Security
Organized by the Responsable AI Team
Important Information
Date: September 17th, 2025
Time: 9h30 - 17h00
Location: Esclangon building, 1st floor, Campus Pierre et Marie Curie, 4 place Jussieu, 75005 Paris
Language: English
Registration: Mandatory (Limited places)
Cost: Free
Description
This workshop focuses on adversarial and backdoor attacks targeting real-life machine learning systems. We will explore vulnerabilities in deployed learning systems, examine attack vectors in practical scenarios, and discuss defense mechanisms for robust ML deployment. The workshop is inspired by research from the KINAITICS project, which investigates kinematic indicators for adversarial behavior detection in AI systems. The event brings together researchers, academics, and industry professionals to discuss cutting-edge developments in adversarial machine learning, security implications, and mitigation strategies for production environments.
Workshop Agenda
September 17th, 2025
9h30 - 9h45: Registration & Welcome Coffee
9h45 - 10h00: Opening Remarks
Rafaël Pinot & Cédric Gouy-Pailler
10h00 - 11h00: Keynote 1: Adversarial attacks and mitigations
Benjamin Negrevergne
11h00 - 11h20: Coffee Break
11h20 - 12h15: Session 1: Real-world Attack Scenarios
[Speaker(s)]
12h15 - 13h45: Lunch Break
13h45 - 14h45: Keynote 2: Backdoors in Artificial Intelligence: Stealth Weapon or Structural Weakness?
Kassem Kallas
14h45 - 15h30: Session 2: Defense Mechanisms & Mitigation
[Speaker(s)]
15h30 - 15h45: Coffee Break
15h45 - 16h30: Session 3: Industry Case Studies
[Speaker(s)]
16h30 - 17h00: Closing Remarks & Networking
Organizers