How much did AlphaGo Zero cost? (2018) • Hacker News Board

Re: How much did AlphaGo Zero cost? (2018)

#141

Post by klipt » Thu, Jun 11, 2020, 6:57 PM UTC

JoeDaDude wrote: post #133

luisvictoria wrote: post #99

Earlier quoted context omitted.

Not really, humans barely provide insight (if anything), which chess engines don’t already consider. Deep Blue could evaluate 200 million different moves... per second. And that’s from 1997. The few and rare times an engine gets funky is usually in end-game positions where the engine can’t seem to find a sacrifice to win the game and will output a current position as drawn. These cases are few and I very much doubt t…

> ...learning completely on its own giving it nothing but the rules which is how AlphaGo works... Not to be too picky, but it was AlphaGo _Zero_ that learned from the rules alone. AlphaGo learned from a large database of human played games: "...trained by a novel combination of supervised learning from human expert games". [1] AlphaGo Zero, derived from AlphaGo, was "an algorithm based solely on reinforcement learnin…

Also AlphaGo Zero never played chess, only go. It was AlphaZero that applied the same framework to other games including chess.

https://en.wikipedia.org/wiki/AlphaGo_Zero https://en.wikipedia.org/wiki/AlphaZero

Re: How much did AlphaGo Zero cost? (2018)

#142

Post by Strilanc » Thu, Jun 11, 2020, 7:35 PM UTC

pps43 wrote: post #97

air7 wrote: post #93

Earlier quoted context omitted.

Is human&computer better than computer only?

Not really. It was tried and it turns out that the best strategy for the human is to just do what the computer suggests.

Could you provide a source for this?

Re: How much did AlphaGo Zero cost? (2018)

#143

Post by hjnilsson » Thu, Jun 11, 2020, 7:40 PM UTC

Scarblac wrote: post #45

hjnilsson wrote: post #7

Another way of thinking about how efficient the brain is: By the article’s numbers, about 5.5 million TPU hours were required to train the machine to play as well as a Go champion. A Go champion might have trained for 8 hours a day, for 15 years (age 5 to 20). That is about 40 000 hours. In other words, machines required 137 times longer to learn the game, and at twice the power consumption! There is still a lot of r…

But there are also many other people spending time studying Go who didn't reach that level. We ran all that studying in parallel and then selected the best person by running a world championship. You can't only count his effort alone.

True. But that single brain, in that person was that efficient. And represents the theoretical gap in efficiency to the machine.

There are for example, other NNs also being trained to play Go, should all unsuccessful attempts be counted into the machine total? The comparison is almost impossible then.

Re: How much did AlphaGo Zero cost? (2018)

#144

Post by hjnilsson » Thu, Jun 11, 2020, 7:42 PM UTC

FartyMcFarter wrote: post #15

hjnilsson wrote: post #7

Another way of thinking about how efficient the brain is: By the article’s numbers, about 5.5 million TPU hours were required to train the machine to play as well as a Go champion. A Go champion might have trained for 8 hours a day, for 15 years (age 5 to 20). That is about 40 000 hours. In other words, machines required 137 times longer to learn the game, and at twice the power consumption! There is still a lot of r…

Go champions don't learn from zero. They learn from teachers, books, and playing against each other. This knowledge is built over hundreds, or thousands of years.

Yes! So perhaps one way to make the machine more efficient, is by one of pre-programmed “general” models, that can be attuned to a particular problem in a much shorter time?

Re: How much did AlphaGo Zero cost? (2018)

#145

Post by ViViDboarder » Thu, Jun 11, 2020, 7:44 PM UTC

sukilot wrote: post #114

tech-historian wrote: post #111

Achieving a new breakthrough in computing is often very expensive. Deep Blue is estimated to cost IBM over $100 million over a decade [1]. And in comparison to large tech company R&D budgets, the amount cited in the article is a drop in the bucket. Consider the fact that Google spent $26 billion in R&D budget in 2019 alone [2]. Microsoft spent almost $17 billion [3]. [1] https://www.extremetech.com/computing/76552-pr…

Note that everything in software development is R&D. Building (not operating) Gmail and Android and Office and Azure are R&D.

Did you mean to say “not everything”? Much may be, but certainly not everything. As you said, operating or maintaining software is not generally considered R&D. Development without the Research component is just Development, not Research and Development.

Re: How much did AlphaGo Zero cost? (2018)

#146

Post by empath75 » Thu, Jun 11, 2020, 8:31 PM UTC

ipsum2 wrote: post #2

Alpha Go Zero*, which was trained from scratch, without human games. I've also heard rumors that AlphaStar ( https://deepmind.com/blog/article/alphastar-mastering-real-t... ) was essentially put on hold because it was too expensive to improve/train. The bot wasn't able to beat StarCraft champions and _only_ got to a grandmaster level.

playing at grandmaster level is a pretty astounding achievement.

Re: How much did AlphaGo Zero cost? (2018)

#147

Post by mehh » Thu, Jun 11, 2020, 8:39 PM UTC

geodel wrote: post #127

sukilot wrote: post #114

Earlier quoted context omitted.

Note that everything in software development is R&D. Building (not operating) Gmail and Android and Office and Azure are R&D.

Right. That's the Development part of Research and Development.

And they get tax breaks, so we are motivated to class as much as possible as R&D.

Re: How much did AlphaGo Zero cost? (2018)

#148

Post by microtherion » Thu, Jun 11, 2020, 8:43 PM UTC

blueboo wrote: post #139

mikorym wrote: post #91

Earlier quoted context omitted.

AFAIK Garry Kasparov to this day does computer&human vs. computer&human chess research, and it's far from a solved problem.

He recently spoke quite dismissively of computer- augmented chess on the Lex Friedman podcast. Essentially, the computer knows best...so computer and human isn’t meaningfully different from computer and rubber stamper.

Maybe the future of computer augmented chess is to form teams of a computer, a person, and a dog. The computer comes up with chess moves. The person feeds the dog. The dog makes sure the person does not touch the board.

Re: How much did AlphaGo Zero cost? (2018)

#149

Post by tuxiano » Thu, Jun 11, 2020, 8:44 PM UTC

Strilanc wrote: post #142

pps43 wrote: post #97

Earlier quoted context omitted.

Not really. It was tried and it turns out that the best strategy for the human is to just do what the computer suggests.

Could you provide a source for this?

I suppose he can’t because it isn’t true at all. The best correspondence players usually improve significantly over the computer suggestions.

Source: I’m a corrspondence chess international master

Re: How much did AlphaGo Zero cost? (2018)

#150

Post by tuxiano » Thu, Jun 11, 2020, 8:49 PM UTC

blueboo wrote: post #139

mikorym wrote: post #91

Earlier quoted context omitted.

AFAIK Garry Kasparov to this day does computer&human vs. computer&human chess research, and it's far from a solved problem.

He recently spoke quite dismissively of computer- augmented chess on the Lex Friedman podcast. Essentially, the computer knows best...so computer and human isn’t meaningfully different from computer and rubber stamper.

I strongly disagree. The best correspondence chess players often improve over the computer suggestions. It takes time, energy and a great strategic knowledge, but it’s still possibile.

Source: I’m a correspondence chess international master