Re:Implementing Neural Turing Machines
by YorkshireSeason on Thursday July 26, 2018 @03:54AM
(#17615807)
Important and unsurprising sentence from the paper's abstract:
"A number
of open source implementations of NTMs exist but are unstable during
training and/or fail to replicate the reported performance of NTMs"
[ Parent
| on HN ]
Why do you deem it unsurprising? Do you think that the opensource neural network implementations are subpar?
Re:Implementing Neural Turing Machines
by YorkshireSeason on Thursday July 26, 2018 @04:17AM
(#17615895)
Anecdotally: I've tried to replicate some recent AI/ML papers and failed. So have some of my acquaintances.
Re:Implementing Neural Turing Machines
by nickpsecurity on Thursday July 26, 2018 @10:12AM
(#17617657)
Some have written about a "reproducibility crisis" in machine learning research:
https://petewarden.com/2018/03/19/the-machine-learning-repro...
Re:Implementing Neural Turing Machines
by albertzeyer on Thursday July 26, 2018 @04:24AM
(#17615916)
I would expect that it was just as unstable for DeepMind. But they can automatically run lots of experiments with different hyper params so that is less an issue for them.