Hacker News Reader

Comment by onra87 | Hacker News Reader

Nice timing ! :)

A few hours ago, Zuckerberg said:

The ancient Chinese game of Go is one of the last games where the best human players can still beat the best artificial intelligence players. Last year, the Facebook AI Research team started creating an AI that can learn to play Go. Scientists have been trying to teach computers to win at Go for 20 years. We're getting close, and in the past six months we've built an AI that can make moves in as fast as 0.1 seconds and still be as good as previous systems that took years to build.

https://www.facebook.com/zuck/posts/10102619979696481

View on HN

So Facebook versus Google is the next sound thing.

I can imagine Google and Facebook having competing armies in the coming robot wars.

Speaking of which, what would a robot war look like? I imagine a large portion of the effort would be to hack/otherwise persuade the enemy robots to switch sides.

If their intelligence is worth anything, they'll meet, setup a common picnic, no humans allowed.

[(Somewhat) Relevant xkcd](https://xkcd.com/1626/).

This is almost exactly the plot of this short story by Philip K. Dick: https://en.wikipedia.org/wiki/The_Defenders_(short_story)

Here's what the US Army Research Lab thinks warfare will look like in 2050. Reading just the Contents is a pretty good TL;DR but basically augmented humans, micro targeting, lots of misinformation, and so much information that decisions have to be automated to the point that humans can only operate "on the loop" instead of "in the loop".

http://www.defenseinnovationmarketplace.mil/resources/ARL_Vi...

...and the Russians will develop electronic warfare that will make all of those advanced systems useless.

Politics. There are obvious strategies (bomb everything, if war still not over, build bigger bombs and GOTO 10), but that sort of demotivates humans and they stop the war.

It'll be interesting (as in the old fake Chinese curse interesting) to know what happens when less democratic power structures wage war. They are still constrained by economics (trade creates value, if they bomb the shit out of someone they might find themselves cut off from trade, see embargoes and sanctions on Russia) and the possibility of an internal struggle (civil war) is always there.

It will always come back to threatening human lives. You can always smash machines against each other, but ultimately it's pointless until the humans themselves are threatened with their lives.

Would be funny to have competitions similar to the boat race between Oxford & Cambridge.

The movie, Colossus: The Forbin Project comes to my mind :)

Note that the game won by AlphaGo against the human player took place in October last year.

"One of the last games".

I submit - League of Legends, any team video game.

I suspect it wouldn't be difficult to program a computer to beat humans at League, but no one has put real effort into it because of anti-cheat and the fact that it would be looked at more as "hacking" than an intellectual challenge.

With over 10 000 hours of Dota [1] under my belt I am fairly certain that even with perfect mechanical skills [2] a strong human player will still beat the A.I.

Even at the very top pro level, matches are constantly being won & lost purely based on the initial hero drafting phase. Calculating the optimal draft is way more difficult than Go. It's probably not an exaggeration to say that the search space scale difference from dota draft to Go is about the same as from Go to tic-tac-toe. Because it's not only about the 100+ different heroes grouped into combinations, but also every possible game that can happen then with those combinations.

Then once the game starts, the A.I. may be able to respond to actions extremely quickly, with perfect precision. But what should the responses be? This is not a simple thing to answer and humans keep taking completely different approaches as our understanding of the game keeps evolving.

Also, the very best dota bots can currently only beat absolute beginners who don't understand the game at all yet. It takes a few hundred games of practice for a human to go beyond the best A.I. currently available.

[1] A game similar to League of Legends

[2] Directly reading from memory and then directly calling gameplay functions, basically a perfect hack, would get you what is known as "mechanical skills". For example being able to last hit.

I understand what you're saying, but I think you fail to realize that a good A.I. is almost the same thing as a better human brain.

You can just train the A.I. by replaying thousands of professional matches, and then let it train by playing billions of matches extremely quickly. It can even play both sides at the same time, and try out millions of different strategies against each other. It doesn't even matter if there's a limitation on how fast it can click and press keys, since every single action will be perfectly optimal.

Not only that, but you don't even need to write a single line of code to tell it about the rules of Dota. Before you start the training, it doesn't even have to know what each key does, or what happens when you move the mouse. Neural networks are capable of learning all of this from scratch, basically by trial and error.

This is not your typical dota bot. Bots are not A.I.s, so this is a whole different ballpark.

I understand what you're describing and I belive that this will be possible in the distant future. I just haven't seen any evidence of this being even close to possible with todays technology.

I think the primary problem is the search space size. I've seen this type of learning work on simple 8 bit games, and it seems we may finally be at the stage to handle Go. However Dota has many orders of magnitude more different possible moves at any given situation. The total search space grows incredibly fast after every move.

Thus, I do think neural nets can eventually learn how to play well, it's just that there's simply not enough memory or processing power to achieve any success right now.

100^10 is pretty small compared to (19*19)^n.

That doesn't take into account the actual in game decisions which are infinite.

My point was that the actual game is what you need to compare - the choice of heroes is superficial compared to the complexity of the game itself, in either case. "Calculating the optimal draft is way more difficult than Go." is simply massively wrong.

The game field in a computer game will be quantized - even if it uses floating-point arithmetic that's still quantization. Conversely a go board can be scaled up or down without losing the feel of the game. If the grids were the same resolution then a given point in time you have many more choices in go because you can play literally anywhere.

"Calculating the optimal draft is way more difficult than Go." is correct because you can't calculate the optimal draft without calculating all the possible games that can happen. Every hero is unique, you can't really preserve anything from the game calculations of another hero.

There's a question of what the AI is doing - makes more sense with an FPS:

1. The AI is a program running on the computer, so the input is the state of the game. This is basically just an aimbot, and would be trivial to do - just make it wander around randomly and headshot everything instantly.

2. The AI is looking at the screen like a human player, and has to parse the screen data (we could even let them have direct pixel input, not a camera). This would be much harder.

For a turn-based game like Go or Chess, the distinction is vague because the CV required to parse a board is fairly trivial and orthogonal to the problem of strategy.

Any game that involves NLP or difficult CV is still not really doable.

Have you ever played Diplomacy?

THERE GOES MY INTELLECTUAL RELEVANCE

Reply on news.ycombinator.com