occlusion dark spots
rajasekharponakala opened this issue ยท 5 comments
rajasekharponakala commented
peikexin9 commented
Sorry @rajasekharponakala, I do not quite remember the exact size. From just the look, I suspect it should be something around 5x5?
rajasekharponakala commented
@peikexin9: Thanks. For example, is the syntax correct?
python gen_diff.py -occl_size 5x5 blackout 1 0.1 10 100 50 0
peikexin9 commented
@rajasekharponakala try something like -occl_size (5,5)?
rajasekharponakala commented
@peikexin9: As the paper discuss on domain-specific constraints, can we consider them as high-dimensional spaces, which are linear in nature and causing adversarial examples?
peikexin9 commented
@rajasekharponakala I would think so. There must be a space linear to these transformations that the NN should be invariant of (while it is actually not thus resulting in the adversarial example).