Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

DeepMind StarCraft II Demonstration [video]

twitch.tv

Post reply on HN
Print view View original

21–30 of 124 posts

Re: DeepMind StarCraft II Demonstration [video]

#21

Post by Buttons840 »

candiodari wrote: post #8
codeflo wrote: post #3

I'm starting to get annoyed with these DeepMind publicity stunts. They don't release any code, and don't let anyone verify their results. Their chess AI beating Stockfish involved some at least a bit questionable setup. And here? I'm a big fan of TLO, but he's currently not even in the top 50 even with his main race. With Protoss, he's clearly making amateur-level mistakes. Why choose this setting to show off your su…

In all of the games Alphastar was slower than TLO in the build order. I agree that TLO could have been faster, but this was definitely not the deciding factor. I think the main factor was Alpha* just had incredibly accurate assessment of exactly what part of it's army it needed to defend/attack and was very careful.

What's AlphaStar's APM. Can it just move every single unit perfectly in a single frame?

Re: DeepMind StarCraft II Demonstration [video]

#22

Post by samfriedman »

candiodari wrote: post #8
codeflo wrote: post #3

I'm starting to get annoyed with these DeepMind publicity stunts. They don't release any code, and don't let anyone verify their results. Their chess AI beating Stockfish involved some at least a bit questionable setup. And here? I'm a big fan of TLO, but he's currently not even in the top 50 even with his main race. With Protoss, he's clearly making amateur-level mistakes. Why choose this setting to show off your su…

In all of the games Alphastar was slower than TLO in the build order. I agree that TLO could have been faster, but this was definitely not the deciding factor. I think the main factor was Alpha* just had incredibly accurate assessment of exactly what part of it's army it needed to defend/attack and was very careful.

The DeepMind folks on stream just mentioned that Alphastar's inference speed is about 50ms from observation to action.

Re: DeepMind StarCraft II Demonstration [video]

#23

Post by gambler »

apetresc wrote: post #18
codeflo wrote: post #3

I'm starting to get annoyed with these DeepMind publicity stunts. They don't release any code, and don't let anyone verify their results. Their chess AI beating Stockfish involved some at least a bit questionable setup. And here? I'm a big fan of TLO, but he's currently not even in the top 50 even with his main race. With Protoss, he's clearly making amateur-level mistakes. Why choose this setting to show off your su…

The fact that the community has been able to use the published paper to replicate all of AlphaZero's results in both chess, go (and a number of other similar combinatorial games) and get similar performance should lay your concerns to rest.

Can you provide some links?

Re: DeepMind StarCraft II Demonstration [video]

#25

Post by soohyung »

knicholes wrote: post #24

I'm having a hard time knowing if it's using a single agent of the five top agents or an ensemble of their five agents.

They select 5 agents and then put up the professional player against a different agent in each game of the series. So it's a single agent, but a different one in each of the five games.

Re: DeepMind StarCraft II Demonstration [video]

#26

Post by aurelwu »

Buttons840 wrote: post #21
candiodari wrote: post #8

Earlier quoted context omitted.

In all of the games Alphastar was slower than TLO in the build order. I agree that TLO could have been faster, but this was definitely not the deciding factor. I think the main factor was Alpha* just had incredibly accurate assessment of exactly what part of it's army it needed to defend/attack and was very careful.

What's AlphaStar's APM. Can it just move every single unit perfectly in a single frame?

it's not perfect but definitely superhuman from what I see. These are demonstration matches and afaik they want to tune the "mechanical" abilities of the agents to be somewhat in the realm of the strongest human players but it's not a trivial thing as not every action is equal and just picking a APM number would make the AI way weaker than a human in some areas (splitting marines, building/morphing lots of units ) and way stronger in other areas (multitasking)

Re: DeepMind StarCraft II Demonstration [video]

#27

Post by saulrh »

knicholes wrote: post #24

I'm having a hard time knowing if it's using a single agent of the five top agents or an ensemble of their five agents.

It sounds like each match in the Bo5 is against a different agent. It's a hair dirty, but I don't think it's too bad because the same kind of random-selection strategy is just as possible "in the wild".

Re: DeepMind StarCraft II Demonstration [video]

#28

Post by superfx »

gambler wrote: post #20
superfx wrote: post #12

Earlier quoted context omitted.

> Their chess AI beating Stockfish involved some at least a bit questionable setup. I believe their eventual Science paper addressed these concerns.

Which paper? How did they address it?

Science paper: http://science.sciencemag.org/content/362/6419/1140

NYT article summarizing some of the issues addressed: https://www.nytimes.com/2018/12/26/science/chess-artificial-...

Re: DeepMind StarCraft II Demonstration [video]

#29

Post by CJefferson »

apetresc wrote: post #18
codeflo wrote: post #3

I'm starting to get annoyed with these DeepMind publicity stunts. They don't release any code, and don't let anyone verify their results. Their chess AI beating Stockfish involved some at least a bit questionable setup. And here? I'm a big fan of TLO, but he's currently not even in the top 50 even with his main race. With Protoss, he's clearly making amateur-level mistakes. Why choose this setting to show off your su…

The fact that the community has been able to use the published paper to replicate all of AlphaZero's results in both chess, go (and a number of other similar combinatorial games) and get similar performance should lay your concerns to rest.

Where?

Re: DeepMind StarCraft II Demonstration [video]

#30

Post by samfriedman »

Buttons840 wrote: post #21
candiodari wrote: post #8

Earlier quoted context omitted.

In all of the games Alphastar was slower than TLO in the build order. I agree that TLO could have been faster, but this was definitely not the deciding factor. I think the main factor was Alpha* just had incredibly accurate assessment of exactly what part of it's army it needed to defend/attack and was very careful.

What's AlphaStar's APM. Can it just move every single unit perfectly in a single frame?

Commentator just mentioned that the APM they've observed from AlphaStar is comparable to pro players.
Post reply on HN