Which AI risk involves crafting inputs to deliberately fool a model?
- Overfitting
- Pooling
- Tokenization
- Adversarial attack
Explanation
Adversarial
attacks use specially crafted inputs to deceive AI models.
Related MCQs
- Single images
- Small spreadsheets
- Extremely large, complex datasets
- Short text files
اس سوال کو وضاحت کے ساتھ پڑھیں
- Virus
- Voltage
- Vector
- Variety
اس سوال کو وضاحت کے ساتھ پڑھیں
- Classification
- Regression
- Backpropagation
- Clustering
اس سوال کو وضاحت کے ساتھ پڑھیں
- Add noise
- Reduce image size
- Encrypt data
- Convert outputs into a probability distribution
اس سوال کو وضاحت کے ساتھ پڑھیں
- Imputation
- Pooling
- Attention mechanism
- Padding
اس سوال کو وضاحت کے ساتھ پڑھیں
