Bài kiểm tra có liên kết hướng dẫn · Cứng Cấp độ
Câu hỏi bảo vệ nghệ sĩ Glaze và Nightshade
Kiểm tra cách kỹ thuật che giấu kiểu Glaze và đầu độc dữ liệu Nightshade sử dụng các nhiễu loạn được tối ưu hóa cũng như lý do tại sao việc nâng cấp và phát hiện có thể loại bỏ các biện pháp bảo vệ này.
Câu hỏi 1 của 8
What threat is Glaze primarily designed to counter?
Xem trước mọi câu hỏi8 câu hỏi
Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.
- What threat is Glaze primarily designed to counter?
- Theft of physical paintings
- Style mimicry through fine-tuning a model on an artist's work
- Screenshots of artwork on social media
- Deletion of artwork from websites
- How does Nightshade differ from Glaze in its goal?
- Nightshade watermarks images for copyright registration
- Nightshade encrypts images so nobody can view them
- Nightshade poisons training data so models learn wrong associations between concepts and images
- Nightshade blocks web crawlers from a website
- What is Glaze's optimization target when it perturbs an artwork?
- A version of the artwork converted into a different style
- A blank white image
- A random image from the internet
- The artist's signature
- What keeps the perturbations from being obvious to human viewers?
- The tools only change image metadata
- The changes are applied only to the image border
- The tools reduce image resolution
- A perceptual constraint, such as an LPIPS budget, limits visible change
- What did a 2024 study by ETH Zurich and Google DeepMind researchers find?
- Glaze makes images impossible to train on under any conditions
- Simple methods like noisy upscaling could remove Glaze-style protections enough for mimicry
- Nightshade works only on printed art
- Perturbations become stronger over time
- Why is protection from these tools not permanent?
- A published image cannot be updated while attack methods and models keep changing
- The software licenses expire after a year
- The perturbations fade from the file over time
- Models are required by law to ignore protected images
- What does transferability mean in the context of these tools?
- Moving an image from one website to another
- Transferring copyright to the tool's developers
- Converting an image between file formats
- A perturbation optimized against one encoder still working on other models
- Why did the Nightshade researchers argue a modest number of poisoned images could affect a model?
- Because every model is trained on only 100 images
- Because poisoned images are copied automatically across the internet
- Because each concept appears in relatively few training images compared with the whole dataset
- Because models trust poisoned images more than clean ones