3 ms·
The author clearly doesn't understand the Szegedy et al. result and isn't really saying anything interesting. Pretty much ALL machine learning image classifiers
by Teodolfo 11y ago
The author clearly doesn't understand the Szegedy et al. result and isn't really saying anything interesting. Pretty much ALL machine learning image classifiers suffer from these pathologies from Szegedy et al. so calling them "deep learning pathologies" is absurdly misleading. Logistic regression has them too as do almost all algorithms that learn to classify images. Human visual systems can also get confused by sensory input that (too us) looks like some object, we just are much better at integrating evidence from other sources to suppress the mistaken neurons.
- adorable 11y agoI guess you are referring to http://cs.nyu.edu/~zaremba/docs/understanding.pdf http://cs.nyu.edu/~zaremba/docs/understanding.pdf For those interested, the papers and discussion are part of this weekly collection of AI-related news and resources: https://aiweekly.curated.co https://aiweekly.curated.co
- Houshalter 11y agoHe's also referring to this follow up paper which attempts to explain the issue and show that it exists in other models too: http://arxiv.org/abs/1412.6572 http://arxiv.org/abs/1412.6572