Awọn Apeere Ọta ati Idanwo Agbara

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

1. Under a specified threat model, what is an adversarial example?
   • An input crafted to cause a model error using changes allowed by that threat model
   • Any ordinary prediction error on unmodified input
   • An input that must contain tiny, imperceptible pixel changes
   • A model update that improves accuracy on every distribution
2. Kini o ṣe iyatọ ikọlu 'apoti-funfun' lati ikọlu 'apoti dudu'?
   • Awọn ikọlu apoti funfun nikan ṣiṣẹ lori awọn aworan funfun
   • Ni funfun-apoti, awọn attacker ni wiwọle si awọn awoṣe ká internals / gradients; ni dudu-apoti, nikan àbájade
   • Black-apoti ku ni o wa nigbagbogbo yiyara
   • Awọn ikọlu apoti funfun ko nilo iṣiro
3. FGSM (Ọna Ami Gradient Sare) ṣe apẹẹrẹ ọta nipasẹ...
   • Awọn piksẹli ti o yipada laileto
   • Retraining awọn awoṣe lati ibere
   • Gbigbe igbesẹ kan ni itọsọna ti ami ti isonu pipadanu pẹlu ọwọ si titẹ sii
   • Gige aworan naa
4. What is adversarial training commonly used for?
   • Guaranteeing robustness to every possible attack
   • Proving a classifier is fair
   • Eliminating all clean-data errors
   • Improving robustness against the adversarial threat model represented during training
5. Kini idi ti 'gbigbe' ti awọn apẹẹrẹ ọta nipa?
   • Ikọlu ti a ṣe lori awoṣe kan nigbagbogbo aṣiwere awọn awoṣe miiran, ti o mu ki awọn ikọlu apoti dudu ṣiṣẹ
   • O mu ki awọn awoṣe reluwe losokepupo
   • O mu awọn idiyele ipamọ pọ si
   • O kan awọn awoṣe ohun nikan
6. How are prompt-injection attacks related to image adversarial examples?
   • They are identical bounded pixel perturbations
   • Both exploit model vulnerabilities, but prompt injection has a distinct instruction and system threat model
   • Prompt injection requires changing model weights
   • Image adversarial examples only affect text prompts
7. Why must an adversarial-robustness report specify the perturbation norm and budget?
   • They determine the model’s number of parameters
   • They guarantee transfer to every attack
   • They define which input changes the reported robustness claim covers
   • They replace the need to state an attacker’s access
8. Which pair of measures helps distinguish clean performance from performance under a specified attack?
   • Training-set size and GPU memory
   • Tokenization rate and context length
   • Model file size and parameter dtype
   • Clean accuracy and robust accuracy under the stated attack setting