No problem beating most humans already. Beating the best human? How much energy goes into creating a civilization that produces that human? Such humans are rare so you ought to count all the others as part of the cost of making the best human as you can't quite make them on demand (though Laslo Polgar might disagree).
I don't get the energy point. The machine has no health care costs and can play 24/7. Doesn't that count for something?
But your wish will come true. Go isn't a special snowflake. If you have an objective metric of success in a formal universe machines always win.
I think he's amortizing the cost to zero for the AI because the marginal cost per additional AI is much lower than the cost to sustain a civilization to churn out and bin human go players.
Your point is not very different from those who say AlphaGo's is really a human victory because human teams built it. Such a distinction based on a historical trace is not a useful one to make. Similarly, the sum total of energy a modern human in a developed society has available to it is not an insightful observation to make when talking about playing Go. It is more a reflection of a civilization's wealth. The best humans from 100, perhaps even 500 years ago would still give almost all modern humans and computers a very hard time. In fact, computers are even more dependent on a technological society (and so more dependent on a large number of humans) than humans are.
The discussion is energy use at play time. For each given second, a certain number of joules are being used to compute a decision. That number as of today, in an unaided match, is independent of civilization's technological state.
That said, AlphaGo has seen a huge (10x?) gain in efficiency according to David Silver. Still far from a human but nonetheless very impressive drop in just a year.
> If you have an objective metric of success in a formal universe machines always win.
More like, machines will eventually win given enough time and effort put forth into making it so by humans. At least, so far.
This is my argument for universal basic income. The true cost of your hyper-effective, top-0.1% employees isn't just their salaries, it's the cost of the entire society that raised them.
How much energy goes into creating a civilization that produces that human?
Did... you really just say that, about a process that requires not only a civilization, but one sufficiently decadent that it can afford to waste resources making silicon that turns burned coal into pointless game victories?
The machine has no health care costs and can play 24/7.
I think by your previous metrics, you should be counting the heathcare costs of the ops folks who run the hardware, and I guess the healthcare costs of their healthcare workers, ad nauseum.
Hey, they provided the human players with a power outlet and let them draw as much power as they wanted. The competition was totally fair.
But seriously, it's possible that AlphaGo is already much more energy efficient than a human player. The main reason it uses tons of energy, is the tree search part of the algorithm. Where it runs hundreds of thousands of simulated games to further analyze every move. This improves it's skill, but only by a little bit. IIRC, the version without tree search beat the full version 25% of the time. Which would still give it a higher elo than Sedol, which only beat it 20% of the time (and AlphaGo has improved since those games.)
Google is also using custom TPUs, which are claimed to be something like an order of magnitude or more energy efficient than GPUs. And computing technology is only getting more energy efficient with time. In principle, transistors moving around a few electrons are vastly more energy efficient than the very wasteful chemical reactions used in brain. We also know how to "sparsify" nets and remove tons unnecessary connections that could reduce computations a lot. But there's generally no point in doing that because it's not faster on normal hardware.
> IIRC, the version without tree search beat the full version 25% of the time.
That would be amazing but it seems hard to believe. Any references?
I found this (which is also impressive):
AlphaGo team then tested the performance of the policy
networks. At each move, they chose the actions that were
predicted by the policy networks to give the highest
likelihood of a win. Using this strategy, each move took
only 3 ms to compute. They tested their best-performing
policy network against Pachi, the strongest open-source
Go program, and which relies on 100,000 simulations of
MCTS at each turn. AlphaGo's policy network won 85% of
the games against Pachi!
>In a similar matchup, AlphaGo running on multiple computers won all 500 games played against other Go programs, and 77% of games played against AlphaGo running on a single computer.
But the full version of AlphaGo that runs on thousands of computers is much stronger than that, so I was mistaken.
Still, the fact that the non-distributed version is so strong even without tree search is pretty amazing. It beat all existing Go playing programs a majority of the time. And with algorithmic advances and more training it may eventually catch up to best human players.
I don't know about the 25% figure, but the original Alpha GO paper mentioned that their best solution is a hybrid approach between neural nets and MCTS. However, the system can beat the best Go bots out there without doing MCTS and relying solemnly on the policy/value networks, which I think is truly amazing.
According to the old paper (AlphaGo has seen significant improvement in efficiency and algorithm so this might be outdated), the distributed version with 1900 CPUs and 280 GPUs defeated the version with 48 CPU and 8 GPUs 81% of the time.
Non-distributed Alpha Go won 99% of the time versus just the value network and policy network with no rollouts. That AI was estimated as having a 2177 Elo rating, which is not very strong and much weaker than Sedol.
Even with a TPU, a human is more efficient. That neural net pair used 8 GPUs. At a generous 200 watt per GPU that's 1.6 kW, 10% of which is 160 watts. A human brain does all higher level reasoning and uses ~20 Watts. A human is not devoting 100% of its computational power on Go. It is likely just a fraction of that.
But if we look at Chess, Chess engines that run on mobile phones are possibly about or maybe slightly more efficient.
>In a similar matchup, AlphaGo running on multiple computers won all 500 games played against other Go programs, and 77% of games played against AlphaGo running on a single computer.
But you are right, the full version running on thousands of computers is much stronger than that.
Still, the fact that the non-distributed version is so strong even without tree search is pretty amazing. With algorithmic advances and more training it may eventually catch up to best human players. It's only the first generation of deep learning based Go bots.
And I believe the policy network only takes a few milliseconds to compute a move. So even if the TPU consumes hundreds of watts at full use, it doesn't need to run at full use for long.
In the post-game press conference, DeepMind mentioned that the version currently playing uses 10x less compute than the Lee Sedol version and runs on a single machine.
That improvement took about decade for computer chess, through a combination of algorithm and hardware improvements. A decade might be a reasonable ballpark estimate for Go as well (but even less than that is realistic). On the hardware side, I would not be surprised if analog circuits make a comeback.
Is that the cumulative energy over the lifetime of the human player (time spent training our neural networks), or just the energy spent during the match?
I don't get the energy point. The machine has no health care costs and can play 24/7. Doesn't that count for something?
But your wish will come true. Go isn't a special snowflake. If you have an objective metric of success in a formal universe machines always win.