Connectionists: Announcing the SaTML’25 Red Teaming Competition on Privacy in Document Visual Question Answering (DocVQA)

[Raouf Kerkouche]

Dear Colleagues,


We are excited to invite you and your colleagues to participate in the Red Teaming Competition at SaTML'25, titled "Inference Attacks Against Document VQA", which focuses on privacy vulnerabilities in Document Visual Question Answering (DocVQA) models. This competition presents a unique opportunity to explore critical privacy concerns in AI, specifically through the development of membership inference and reconstruction attacks tailored to DocVQA models.


We would greatly appreciate it if you could share this announcement with anyone in your network who might be interested in participating.


More details, including competition rules and registration, can be found on the official website: https://benchmarks.elsa-ai.eu/?ch=2&com=introduction.

Important Dates:

  *   Submission Deadline: February 28, 2025
  *   Announcement of Winning Teams: March 10, 2025
  *   Competition Track: April 9-11, 2025 (during SaTML’25)

For any inquiries or further information, please do not hesitate to contact us at info_pfl at cvc.uab.cat.


We look forward to your participation and the innovative solutions you will bring to this critical area of AI privacy research.


The competition is organized by the European Lighthouse on Secure and Safe AI (ELSA).


Best regards,


Dr. Raouf KERKOUCHE

Postdoctoral Researcher,

CISPA Helmholtz Center for Information Security


Stuhlsatzenhaus 5, Saarland Informatics Campus
66123 Saarbrücken, Germany

________________________________
From: Raouf Kerkouche
Sent: Thursday, November 14, 2024 9:49:59 AM
To: connectionists at mailman.srv.cs.cmu.edu
Subject: Announcing the SaTML’25 Red Teaming Competition on Privacy in Document Visual Question Answering (DocVQA)


Dear Colleagues,


We are excited to invite you and your colleagues to participate in the Red Teaming Competition at SaTML'25, titled "Inference Attacks Against Document VQA", which focuses on privacy vulnerabilities in Document Visual Question Answering (DocVQA) models. This competition presents a unique opportunity to explore critical privacy concerns in AI, specifically through the development of membership inference and reconstruction attacks tailored to DocVQA models.


We would greatly appreciate it if you could share this announcement with anyone in your network who might be interested in participating.


More details, including competition rules and registration, can be found on the official website: https://benchmarks.elsa-ai.eu/?ch=2&com=introduction.

Important Dates:

  *   Submission Deadline: February 28, 2025
  *   Announcement of Winning Teams: March 10, 2025
  *   Competition Track: April 9-11, 2025 (during SaTML’25)

For any inquiries or further information, please do not hesitate to contact us at info_pfl at cvc.uab.cat.


We look forward to your participation and the innovative solutions you will bring to this critical area of AI privacy research.


The competition is organized by the European Lighthouse on Secure and Safe AI (ELSA).


Best regards,


Dr. Raouf KERKOUCHE

Postdoctoral Researcher,

CISPA Helmholtz Center for Information Security


Stuhlsatzenhaus 5, Saarland Informatics Campus
66123 Saarbrücken, Germany
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20241114/ebe95f28/attachment.html>