privacy risks
Privacy risks in AI relate to potential threats to individual privacy or data security that arise from using AI systems, particularly in terms of data misuse, unauthorized access, or the inability to ensure confidentiality in model training and inference.
- Differentially Private Relational Learning with Entity-level Privacy Guarantees
- LLM Meeting Decision Trees on Tabular Data
- Perturb a Model, Not an Image: Towards Robust Privacy Protection via Anti-Personalized Diffusion Models
- Self-Refining Language Model Anonymizers via Adversarial Distillation
- Setting $\varepsilon$ is not the Issue in Differential Privacy
- Subgraph Federated Learning via Spectral Methods
- Unlearned but Not Forgotten: Data Extraction after Exact Unlearning in LLM
- VMDT: Decoding the Trustworthiness of Video Foundation Models