Hacker News Reader

Comment by amelius | Hacker News Reader

I think NVidia deserves part of the award, as they have been a major factor in the progress of NNs.

Is this a meme? No they have not. Sure they are a popular hardware provider, but NN have been around before they were even a company. I think Fukushima deserves a spot here too, before giving it to Nvidia, but there are so many people that have contributed to NN, so I guess those in the spotlight take the glory if it must be taken.

...Fukushima? Is it a typo for someone (or some company)?

He's credited with first introducing the CNN architecture (Neocognitron) and a way to train it with unsupervised learning, inspired by the neurological research of Hubel and Wiesel and his earlier work in training neural networks with unsupervised learning (backprop is now used instead). Probably one of the most famous neural network papers. https://www.rctn.org/bruno/public/papers/Fukushima1980.pdf

http://vision.stanford.edu/teaching/cs131_fall1415/lectures/...

For making insanely painful-to-deal with dependencies and effectively instituting de facto vendor lock-in for deep learning?

Nah.

They don't deserve science award because they didn't do science or pioneer anything. They just responded to demand. They have been awarded with money.

It was the high performance scientific computing community that started to use graphics cards to perform matrix computations in the early 2000's. The term GPGPU term was invented. GPU programming pain in the ass in early 2000's. Nvidia saw some markets in GPGPU's for scientific computing. First release of CUDA was 2007.

> They don't deserve science award because they didn't do science or pioneer anything.

NVidia didn't do science? It depends on what you call science (math and computer science are not "science" according to a very strict definition involving falsifiability). But in any case, an enormous amount of research went into the development of GPUs. It's not just engineering work. Lots of PhD (students) contributed.

the context is DL.

True, but neural networks existed for a long time. The hardware is what made NNs actually practical. Recent theoretical advances have been only incremental. For example, CNNs are just evolutionary since image recognition has used convolutions long before NNs were practical, and CNNs are just a straightforward translation of existing techniques to NNs.

Sure theoretical advances may be evolutionary, but they resulted in exponential reduction of parameter space and sample complexity. These advances outpace Moore's law by a large margin.

Hardware served as a catalyst, but it was not a necessity.

good point, I think awards should go to NVidia

Could you be more specific? I was inclined to respond, "Intel, too. And keyboard and monitor manufacturers. And throw in John Von Neumann!"

Von Neumann thought about artificial neural networks too: https://ia800800.us.archive.org/4/items/TheComputerAndTheBra...

By that same logic Google deserves part of Vint Cerf's Turing Award.

Please , lets not buy into ever marketing hype. AFAIK GPUs were invented for driving virtual cars

Reply on news.ycombinator.com