3 ms·
Sorry what's your contention? or do you have one? Backprop is backprop on CPU or GPU - it just runs (way) faster on GPU. So while you can do some stuff with NN
by ShamelessC 4y ago
Sorry what's your contention? or do you have one?
Backprop is backprop on CPU or GPU - it just runs (way) faster on GPU. So while you can do some stuff with NN's on CPU, you can do more stuff with NN's on GPU.
Further, the sources you reference are from before advances made in software and hardware for massively parallel GPGPU architectures (CUDA), so it's kind of a moot point anyways. Those researchers may have had access to such a system, but you definitely couldn't just `pip install pytorch` on your gaming GPU in the 90's...
- deepnotderp 4y agoDan Ciresan’s DanNet ran on GPUs before Hinton’s students
- P-NP 4y agoSchmidhuber's DanNet pages: https://people.idsia.ch/~juergen/DanNet-triggers-deep-CNN-revolution-2011.html https://people.idsia.ch/~juergen/DanNet-triggers-deep-CNN-re... https://people.idsia.ch/~juergen/2010-breakthrough-supervised-deep-learning.html https://people.idsia.ch/~juergen/2010-breakthrough-supervise...
- ekleraki 4y ago> Sorry what's your contention? Time of publication. I merely pointed that Schmidhuber et al's results were published long before Hinton et al's work (I included year of publication for a reason), ergo the comment that was in support of Hinton was refuted and my comment gives further validity to Schmidhuber's claims of misattribution and plagiarism, while pointing at yet another paper tiger in the claims that the three pioneered the field. If anything, Schmidhuber may have deserved a Turing award on his own, and what we see is the result of politics. This is pure speculation, but to me the whole ordeal hints at manipulation from the companies where 2 of the 3 worked at, which would essentially increase the pool of candidates the respective companies have only by association with their employee.
- ShamelessC 4y agoAh, that makes sense. Thanks for clarifying.