News for hackers. Stuff that matters.

A comment on How convolutional neural networks see the world.

Re:How convolutional neural networks see the world by mdonahoe on Sunday January 31, 2016 @09:21PM (#11009269)
What would happen if you generated a lot of images like this, tagged them as "conv-net filter", and then added them to the set of images that the neural nets trained on?

Would the network learn more discriminating filters for everything else?

[ Parent | on HN ]
Re:How convolutional neural networks see the world by dhj on Sunday January 31, 2016 @11:19PM (#11009746)
Very interesting question and worth an experiment or two. Personally, I don't think it would help because the patterns are complex and it is an internal representation. It would be like feeding it a hash values and labelling them "hash values". That wouldn't necessarily help you discriminate pre-hashed values. That said, I would definitely be interested in the results of a real test.
[ Parent | on HN ]