3 ms·
> Very, very easy to adjust the algorithm for a t-shirt. The operative point here is not 'a shirt', but a visual pattern that tricks deep learning-style classi
by crooked-v 7y ago
> Very, very easy to adjust the algorithm for a t-shirt.
The operative point here is not 'a shirt', but a visual pattern that tricks deep learning-style classifiers into wildly misidentifying something. There's no 'very easy' way to counteract that other than retraining on a new dataset or switching entirely away from a deep learning system.
- dTal 7y agoSurely all you have to do is train a "trippy shirt" detector, and run it in parallel?
- Kalium 7y agoAs I understand it, adversarial designs generally work on one specific recognition system. So working around this attack would be very achievable with three or more recognition systems and a consensus check. This particular paper is based around attacking YOLOv2.
- polishTar 7y agoI think these types of adversarial attacks are even easier to foil than that because they're specific to one particular set of weights. Even really really small changes in the training data or model could invalidate the attack if I understand correctly.
- andrewflnr 7y agoI know there has been work in generating adversarial images that work against multiple models. That kind of thing is probably only going to get better, to say nothing of particular sets of weights in a single model.