5 ms·

This is why about 90% of the papers Google publishes are useless. They describe deep learning successes, with datasets that aren't public, on hardware that isn

This is why about 90% of the papers Google publishes are useless. They describe deep learning successes, with datasets that aren't public, on hardware that isn't public, with software that isn't public. They describe technologies, be it spanner, borg, flumejava or others, relying on closed hard- and software. If an average university professor can't replicate research 1:1 based solely on existing, open technology and the paper, the paper should not be accepted by any journal. In fact, it shouldn't even be accepted on conferences, or be able to get a doi code. That's not worthy of being called science.

remus 9y ago
I don't think they're quite useless (they tend to contain some interesting ideas, which presumably work as google have generally implemented them) but as you say they're definitely not 'science' as they can't be replicated.
dagw 9y ago
While I obviously agree that fully reproducible results is the ideal, just knowing that something is possible and having some idea where to start is often valuable in itself.
- kuschku 9y ago
  First of all, if it’s not reproducible, it’s not science. The entire point of science is getting results that give insight to others. Second, often the papers set prerequisites that are not available to anyone else, because these, too, are undocumented. This is cool if you want an inspiration, but large parts of these papers are as useful as watching Star Trek for actual science. They can serve as inspiration, but they can’t tell you how to replicate something, or give you possible leads to expand on it. And this issue gets even worse – if I’m writing an undergrad thesis on a tech that Google has several years lead on, my thesis won’t be able to produce anything new. I can’t expand on the state of the art, because the state of the art isn’t even available to me.

remus 9y ago

I don't think they're quite useless (they tend to contain some interesting ideas, which presumably work as google have generally implemented them) but as you say they're definitely not 'science' as they can't be replicated.

dagw 9y ago

While I obviously agree that fully reproducible results is the ideal, just knowing that something is possible and having some idea where to start is often valuable in itself.

kuschku 9y ago
First of all, if it’s not reproducible, it’s not science. The entire point of science is getting results that give insight to others. Second, often the papers set prerequisites that are not available to anyone else, because these, too, are undocumented. This is cool if you want an inspiration, but large parts of these papers are as useful as watching Star Trek for actual science. They can serve as inspiration, but they can’t tell you how to replicate something, or give you possible leads to expand on it. And this issue gets even worse – if I’m writing an undergrad thesis on a tech that Google has several years lead on, my thesis won’t be able to produce anything new. I can’t expand on the state of the art, because the state of the art isn’t even available to me.