News for hackers. Stuff that matters.
Re:A way to deal with enormous branching factors in strategy... by tschellenbach on Friday March 25, 2016 @02:18PM (#11361728)
I wonder how long it will take till we find a way to accelerate the rate of learning. If you give a user (experienced with games), a tablet with the FTL game, chances are they will beat it in 10-30 attempts. AI on the other hand still takes thousands of playthroughs to get good at a game.
Re:A way to deal with enormous branching factors in strategy... by ivanhoe on Friday March 25, 2016 @03:07PM (#11362001)
Well, to be fair even the youngest gamer has spent years and years as baby/toddler focused almost solely on learning how to use his/her brain to reason and make quick strategic decisions. As any parent knows babies know almost nothing and have to learn it all from the scratch and watching them figuring out the world and improving their skills is very insightful to anyone interested in AI. And we often forget how much time and hard work we all invested as kids to perfect our (basic) abilities...
Re:A way to deal with enormous branching factors in strategy... by Houshalter on Friday March 25, 2016 @07:02PM (#11363396)
A good example is when you throw humans into a game with physics they aren't familiar with. E.g a 4 dimensional game. Or QWOP. Humans then suddenly don't seem so amazing without also taking thousands of attempts.
Re:A way to deal with enormous branching factors in strategy... by infogulch on Friday March 25, 2016 @04:05PM (#11362292)
I feel like the ways humans and AI are trained is analogous to vertical / horizontal integration, respectively. I can't quite describe why I feel this.
Re:A way to deal with enormous branching factors in strategy... by mey on Friday March 25, 2016 @02:28PM (#11361783)
An AI is typically starting with no prior experience and ability to reason. It is essentially brute forcing approaches and is basically succeeding based on the birthday paradox.

https://en.wikipedia.org/wiki/Hacker_koan#Uncarved_block

Re:A way to deal with enormous branching factors in strategy... by namenotrequired on Friday March 25, 2016 @03:46PM (#11362202)
The funny thing about that link is that the original anecdote is far more enlightening than the "koan" it was turned into.
Re:A way to deal with enormous branching factors in strategy... by GFK_of_xmaspast on Friday March 25, 2016 @03:29PM (#11362120)
What does the birthday paradox have to do with this.
Re:A way to deal with enormous branching factors in strategy... by elbigbad on Friday March 25, 2016 @03:33PM (#11362140)
I assume because it would take a fewer number attempts than one would expect intuitively before the AI finds a successful strategy. Thus severely cutting down on the amount of time (or number of iterations) before a successful strategy is found.

Like with the birthday paradox, one would intuitively think it might take hundreds of people in a room before two had the same birthday, but of course we know that is indeed not the case.

Re:A way to deal with enormous branching factors in strategy... by schoen on Friday March 25, 2016 @04:44PM (#11362553)
The birthday paradox involves the possibility of random processes coinciding at some point, which feels a bit different from what the AI is doing.

Birthday paradox: different people are coming into the room with different randomly-distributed birthdays; how many people do you expect need to come in before two people in the room will have the same birthday?

Ordinary random search: there are several people in the room with birthdays that are unknown to you. How many birthdays do you expect you'll have to guess before you name a birthday that is actually the birthday of someone present?

Re:A way to deal with enormous branching factors in strategy... by pavel_lishin on Friday March 25, 2016 @06:04PM (#11363032)
I've played it probably fifty times, and I've only gotten to the boss level three times.

But it's possible that I'm just bad at video gaes.

Re:A way to deal with enormous branching factors in strategy... by simoncion on Friday March 25, 2016 @08:52PM (#11363847)
Are you playing on Easy or Normal? If you've that result while playing on Normal, then you're not bad at video games.
Re:A way to deal with enormous branching factors in strategy... by talmand on Friday March 25, 2016 @03:24PM (#11362094)
Ha! On easy maybe.