adversarial deep learning
Adversarial deep learning is a technique in machine learning where a neural network is trained using a two-player game-like framework. One player generates realistic inputs, known as adversarial examples, to deceive the network, while the other player tries to detect and classify them correctly. This process helps improve the network's robustness by forcing it to learn from challenging and contrary examples.
Requires login.
Related Concepts (1)
Similar Concepts
- adversarial anomaly detection
- adversarial attacks
- adversarial autoencoders
- adversarial detection and defense
- adversarial examples
- adversarial examples and attacks on ai systems
- adversarial feature learning
- adversarial image classification
- adversarial image synthesis
- adversarial input synthesis
- adversarial machine learning
- adversarial reinforcement learning
- adversarial robustness
- adversarial text generation
- adversarial training