Towards Understanding and Enhancing Robustness of Deep Learning Models against Malicious Unlearning Attacks.
Wei QianChenxu ZhaoWei LeMeiyi MaMengdi HuaiPublished in: KDD (2023)
Keyphrases
- learning models
- semi fragile watermarking
- malicious users
- detect malicious
- malicious attacks
- loss function
- machine learning
- learning tasks
- detecting malicious
- semi supervised learning
- conditional random fields
- machine learning algorithms
- learning algorithm
- learning problems
- classification models
- machine learning models
- decision trees
- sparse metric learning
- genetic algorithm
- watermarking scheme
- higher order
- supervised learning
- pairwise