Corrupting training data to manipulate model behavior or introduce vulnerabilities.
This concept is essential for understanding specialized ai topics and forms a key part of modern AI systems.
Related Concepts
- Security
- Backdoor Attack
- Training
Corrupting training data to manipulate model behavior or introduce vulnerabilities.
A collection of data examples used for training, validating, or testing machine learning models.
The examples a model learns its weights from, kept separate from the validation and test data used to check how well it generalizes.
Building systems that learn patterns from data instead of following hand-written rules, getting better at a task as they see more examples.
Protecting machine learning systems from attackers who tamper with training data, craft inputs that fool the model, or extract private data or the model itself.
Corrupting training data to manipulate model behavior or introduce vulnerabilities.
This concept is essential for understanding specialized ai topics and forms a key part of modern AI systems.
A paper that builds on Data Poisoning .