lerid

4 ms·

1. 1.▲
   by lerid 10y ago

   These adversarial inputs have been shown to generalize to separately trained models.
2. 2.▲
   by lerid 10y ago

   This seems like an obvious thing to try, but technically the convolutional structure already looks at different scales (in a way similar to mipmaps).
3. 3.▲
   by lerid 10y ago

   These adversarial inputs have been shown to generalize to separately trained models.
4. 4.▲
   by lerid 10y ago

   Did you read the article? > We find that both adversarial training and defensive distillation accidentally perform a kind of gradient masking. Neither algorithm was explicitly designed to perform gradient masking, but gradient masking is
5. 5.▲
   Ask HN: Google no longer hiring Software Engineers in Zurich?

   3 points by lerid 10y ago|0 comments