3 ms·

There's some great stuff there, but that list could use an update. The most recent papers seem to be from 2011. Things move very fast in this field. This was ba

by benanne 12y ago

There's some great stuff there, but that list could use an update. The most recent papers seem to be from 2011. Things move very fast in this field. This was back when pre-training was still strongly recommended and before Krizhevsky's ImageNet win (2012), which shifted the focus of the field away from unsupervised learning somewhat. The section on convolutional neural networks especially is a bit meager by today's standards :)

jcr 12y ago
You made some really great points, but if Geoffrey Hintons's list of "relevant literature" has nothing newer than 2007, then, it seems nobody has solved the continuous update problem for fast moving fields. ;) http://www.cs.toronto.edu/~hinton/deeprefs.html http://www.cs.toronto.edu/~hinton/deeprefs.html A bit more seriously, the Stanford UFLDL wiki seems to be intended more as an initial tutorial. It's related their CS classes so it provides recommended reading on tutorial-relevant and class-relevant works. It may lack some of the more recent advancements, but it still seems to be an active work in progress, and it's received about 1000 edits in the last 90 days. http://deeplearning.stanford.edu/wiki/index.php/Special:ActiveUsers http://deeplearning.stanford.edu/wiki/index.php/Special:Acti... It's Stanford and it's class/research related, so I'm not sure if the wiki is open to public editing (I haven't tried yet). If you have ideas/links to improve the reading list, I'm sure they'd appreciate either feedback or wiki edits. I think I managed to track down the Krizhevsky paper you mentioned? http://books.nips.cc/papers/files/nips25/NIPS2012_0534.pdf http://books.nips.cc/papers/files/nips25/NIPS2012_0534.pdf Krizhevsky, A., Sutskever, I. and Hinton, G. E. ImageNet Classification with Deep Convolutional Neural Networks NIPS 2012: Neural Information Processing Systems, Lake Tahoe, Nevada Thanks for mentioning it. As a useful resource, the Stanford UFLDL reading list still seemed worth posting to HN. There are just mountains of AI/ML/DL/NN/... papers, projects and related resources, so picking out interesting and useful things for HN can be really difficult.
- benanne 12y ago
  True, there is no real up to date curated reading list that I know of. Actually, there was a list of 2014 deep learning papers going around recently, but it seemed to list basically every tangentially related paper that the authors could find, so it was pretty useless as an introduction to the field.
  - Houshalter 12y ago
    There are some survey papers like this: http://arxiv.org/abs/1404.7828 http://arxiv.org/abs/1404.7828 The point isn't to read every paper it references, but to just get a summary of what the current state of research is. If you are interested in an area then you can go read the papers there.

by benanne 12y ago

jcr 12y ago

You made some really great points, but if Geoffrey Hintons's list of "relevant literature" has nothing newer than 2007, then, it seems nobody has solved the continuous update problem for fast moving fields. ;) http://www.cs.toronto.edu/~hinton/deeprefs.html http://www.cs.toronto.edu/~hinton/deeprefs.html A bit more seriously, the Stanford UFLDL wiki seems to be intended more as an initial tutorial. It's related their CS classes so it provides recommended reading on tutorial-relevant and class-relevant works. It may lack some of the more recent advancements, but it still seems to be an active work in progress, and it's received about 1000 edits in the last 90 days. http://deeplearning.stanford.edu/wiki/index.php/Special:ActiveUsers http://deeplearning.stanford.edu/wiki/index.php/Special:Acti... It's Stanford and it's class/research related, so I'm not sure if the wiki is open to public editing (I haven't tried yet). If you have ideas/links to improve the reading list, I'm sure they'd appreciate either feedback or wiki edits. I think I managed to track down the Krizhevsky paper you mentioned? http://books.nips.cc/papers/files/nips25/NIPS2012_0534.pdf http://books.nips.cc/papers/files/nips25/NIPS2012_0534.pdf Krizhevsky, A., Sutskever, I. and Hinton, G. E. ImageNet Classification with Deep Convolutional Neural Networks NIPS 2012: Neural Information Processing Systems, Lake Tahoe, Nevada Thanks for mentioning it. As a useful resource, the Stanford UFLDL reading list still seemed worth posting to HN. There are just mountains of AI/ML/DL/NN/... papers, projects and related resources, so picking out interesting and useful things for HN can be really difficult.

benanne 12y ago
True, there is no real up to date curated reading list that I know of. Actually, there was a list of 2014 deep learning papers going around recently, but it seemed to list basically every tangentially related paper that the authors could find, so it was pretty useless as an introduction to the field.
- Houshalter 12y ago
  There are some survey papers like this: http://arxiv.org/abs/1404.7828 http://arxiv.org/abs/1404.7828 The point isn't to read every paper it references, but to just get a summary of what the current state of research is. If you are interested in an area then you can go read the papers there.