But this has its cons you know. Lot of the work these days are just improving state of the art. IDGAF about sota man. New ideas should flow not architecture tweaking lol
Sometimes I wonder why is the top-5 image classification task so difficult. If you are giving me 5 chances to look at an image and correctly classify it from ~1000 Imagenet classes, I can surely do better than 5-10% error rate.
Also, now that the top-5 error rate been brought down considerably, what is the next benchmark for the research community to beat? A new dataset, top-1 error rate on Imagenet?
A large majority of human errors come from fine-grained categories(such as correctly identifying two similar cat species) and class unawareness. I would recommend this article by Andrej Karpathy, where he talks about his learning from competing against GoogLeNet: http://karpathy.github.io/2014/09/02/what-i-learned-from-com...
There is a point of view out there that Europe's higher than average reliance on renewables has bumped up electricity prices there and contributed in making the place less competitive for industries. You can see that argument in action when European miners are losing out to others due to high power costs.
Generally speaking, any slowdown in economic activity will lead to a slowdown in global warming. The question is always: is the new equilibrium better or worse for all involved? (Think of the unemployed)
I think a better metric would be $GDP / ton CO2 to see who's efficient and who's not. Unfortunately, GDP numbers are not really standardized, at least countrys' individual way of calculation GDP varies greatly.
That point of view seems rather uninformed, because industrial and residential pricing of electricity, gas etc. are virtually unrelated, and the former doesn't include new extra taxes for renewables in any country I know.
Yeah, but it seems manufacturing is not that popular in AI research. Maybe this is because getting the desired data and performing the entire research process is much slower for this domain.
