Empathy-Based Sandbox Approach to Bridge Privacy Gap
An empathy-based approach using artificial personas in a sandbox environment bridges the privacy gap by allowing users to experience and understand the impact of their privacy data on system outcomes.
Private Aggregate Queries to Untrusted Databases: Ensuring Privacy in Data Retrieval
Ensuring privacy in aggregate queries through innovative IT-PIR frameworks.
Less is More: Revisiting the Gaussian Mechanism for Differential Privacy
全ての既存のガウス機構は、完全ランク共分散行列の呪いに苦しんでおり、新しいRank-1特異多変量ガウス（R1SMG）メカニズムはこの呪いを解消します。
Budget Recycling Differential Privacy Framework for Enhanced Utility and Privacy Protection
Introducing the Budget Recycling Differential Privacy (BR-DP) framework enhances utility and privacy protection by optimizing budget allocation and incorporating subsampling.
Differential Privacy Lower Bounds for Continual Observation and Online Threshold Queries
The article presents new lower bounds for differential privacy under continual observation and online threshold queries, challenging existing upper bounds.
Instance-Optimal Clipping for Summation Problems in Shuffle Model of Differential Privacy
Clipping mechanism optimizes error bounds in shuffle-DP for sum estimation problems.
Revisiting the Gaussian Mechanism for Differential Privacy: Less is More
R1SMG mechanism achieves (ε,δ)-DP with lower noise magnitude.
Advancing Differential Privacy: Current Practices and Future Directions
Advancing differential privacy practices for real-world applications.
Investigating Visual Privacy Auditing with Diffusion Models
The author explores the impact of real-world data priors on data reconstruction attacks, highlighting a discrepancy between theoretical models and practical outcomes. The study emphasizes the significance of incorporating data priors accurately into privacy guarantees for better alignment with real-world scenarios.
CypherTalk: Cost-Effective LLM Privacy Mechanism
The author introduces CypherTalk, a cost-effective and self-adaptive shaking and recovery mechanism for Large Language Models (LLMs) to balance privacy concerns with operational efficacy.
© 2024 by Linnk AI