An adversary adds imperceptible noise to evade a facial-recognition system. This is an:
- Data backup
- Adversarial perturbation attack
- OCR error
- Routine update
Explanation
Explanation:
Imperceptible
noise crafted to fool models is an adversarial perturbation attack.
Related MCQs
- k-fold cross-validation
- Single holdout only
- Manual guessing
- No validation
اس سوال کو وضاحت کے ساتھ پڑھیں
- Data poisoning
- Tokenization
- Pooling
- Padding
اس سوال کو وضاحت کے ساتھ پڑھیں
- AI-text detection/classifiers
- Image sharpening
- Disk cloning
- Path planning
اس سوال کو وضاحت کے ساتھ پڑھیں
- They map data into dense numeric vectors capturing meaning
- They delete data
- They are image filters
- They are encryption keys
اس سوال کو وضاحت کے ساتھ پڑھیں
- ANPR using computer vision + OCR
- Text summarization
- NLP sentiment analysis
- Audio clustering
اس سوال کو وضاحت کے ساتھ پڑھیں
