Â鶹´«Ã½AV

Protect-IT focuses on the security and privacy of machine learning (ML) from an information-theoretic standpoint. ML systems rely on extensive datasets, often containing sensitive personal information, posing a threat to user privacy and security. Protect-IT aims to bring together experts in machine learning, (algorithmic) fairness and information theory to explore, develop, and evaluate privacy, security, and fairness attacks against ML algorithms along with defense strategies to counter them. We welcome submissions on the following topics:

• Information leakage, data correlation 
• Defining and quantifying privacy
• Differential privacy and private data analysis
• Machine learning and privacy
• Transparency, robustness and abuse in privacy systems
• Information theoretic foundations of security and privacy
• Cryptographic tools for privacy
• Algorithmic fairness 
• Adversarial learning 
• Intrusion detection
• Privacy and security attacks and defenses

Submission Instructions: We welcome two types of submissions: full papers of maximum 12 pages or extended abstracts up to 4 pages. Full papers must be unpublished elsewhere and will be published in ACM digital library as workshop proceedings. We encourage submissions of work that is new to the community of data privacy, security and information theory. Extended abstracts can be previously published or currently under review elsewhere and will not be included in proceedings. All submissions must adhere the latest ACM Sigconf style conference template. We also welcome recently published studies as well as unpublished recent results in privacy and security venues only as poster submissions.

Poster Instructions: In addition to the accepted poster submissions of recent results, all accepted full paper submissions will have a slot at the poster session for posters with size no bigger than A0 (841 × 1189 mm).

Submission deadline: February 21, 2025, 23:59 (Anywhere on Earth)
Notification of acceptance: April 08, 2025