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

AI Beats Four Top Poker Players

bbc.co.uk

Post reply on HN
Print view View original

81–90 of 234 posts

Re: AI Beats Four Top Poker Players

#81

Post by algorias »

jasonwatkinspdx wrote: post #19
MaggieL wrote: post #3

Surely nobody is surprised? Perfect knowledge of probabilities, no emotion and immune to bluffing.

> immune to bluffing There's no such thing. An algorithm playing straight hand value based on probabilities is more susceptible to bluffing, not less. And this is no limit, where a single hand can swing all the chips. Any poker AI that isn't a loser is going to have some pretty sophisticated modeling of the opponent.

Having a poker AI that plays deterministic strategies is an obviously terrible idea, for the precise reason you mention.

It makes much more sense for the strategy space to be the set of probability distributions over game moves (i.e. mixed strategies).

I think that the optimal mixed strategy for each hand is immune to bluffing (over many hands it will have larger expected winnings against a bluffer). If that wasn't the case, there would exist no Bayes-Nash equilibrium for the game, contradicting Nash's theorem.

Re: AI Beats Four Top Poker Players

#82

Post by brilliantcode »

ori_b wrote: post #76
brilliantcode wrote: post #73

Earlier quoted context omitted.

Sure it is. Much like most developers don't spend their time reading on computing science papers to do their job, I question why such academic rigour is to be demanded from somebody trying to capitalize on the arbitrage opportunity. Math is good but my time won't be best used if I have to learn calculus all over again just to begin understanding the linguo. Rather have a generic model of what to use and when, hire th…

Because this is the language of the domain. Much like code is the language of programming. This isn't academic rigor or lingo. This is how things work. To me, your comment reads similarly to "Much like most business people don't spend their time reading code to do the job, I question why such academic rigor is to be demanded by someone who is translating business requirements into database operations".

I think we are after different things. My goal is not to understand the underlying technology at the academic level but just enough to make money by utilizing the people that do, just like any good capitalist.

You don't need to be an oil rig worker to understand where the opportunity lies is what I'm saying. It's possible to understand systems from a blackbox point of view and still be able to exploit them. Afterall, this is what we do whenever we hire a new developer, it's impossible to know what they know so I leave that up to smarter and talented people while I allocate the capitals and set a vision for what's required to capitalize on the business opportunity that AI presents with it's superior performance, scalability and economies to scale.

You outsource what you don't know or can't do and if you can't afford it you do end up doing it yourself. This is why capitalism works. It's not necessary for you to dwindle with all the underlying complexities when you model the business opportunity in terms of basic expenditure and investment returns.

Re: AI Beats Four Top Poker Players

#83

Post by deepnotderp »

brilliantcode wrote: post #73
deepnotderp wrote: post #44

Earlier quoted context omitted.

> "So the question is, how does an average joe hacker like me exploit and leverage this wonderful thing called deep learning? I'm not interested in reading PHD papers with advanced calculus." FFS, if you're not willing to read a paper with BASIC CALCULUS (it's HS/college, not advanced like fractional), then I'm not sure machine learning is the right place for you.

Sure it is. Much like most developers don't spend their time reading on computing science papers to do their job, I question why such academic rigour is to be demanded from somebody trying to capitalize on the arbitrage opportunity. Math is good but my time won't be best used if I have to learn calculus all over again just to begin understanding the linguo. Rather have a generic model of what to use and when, hire th…

> "...academic rigor..." "...calculus.."

Dude, I'm not even asking you for anything that difficult here, how would you like it if I walked in and said "well, I don't need to learn any programing, my time is better spent magically making it work".

Re: AI Beats Four Top Poker Players

#84

Post by DennisP »

andrewprock wrote: post #56
doesnotexist wrote: post #2

Andrew Ng posted a comment about this on facebook: "I'm thrilled about Libratus' Poker triumph--this is a huge step for AI. Othello/Checkers/Chess/Go were theoretically solvable with minimax tree search and sheer computation; but poker, which requires bluffing, needs sophisticated modeling of your opponents and new algorithmic principles. CMU's Tuomas Sandholm has also (in a private email) promised to publish their a…

I'm surprised that Andrew Ng made this claim. The strategy that was built for Libratus' predecessor did not do sophisticated modeling of the opponents, or use new algorithmic principles. Poker is solved using a very large game tree, just as with the other games. The structure of the tree is modified to support the notion of hidden state, but beyond that it is essentially the same as the other games. The structure for…

The no-limit hold'em game tree has over 10^160 states.

Even limit poker is too large to solve directly. In 2015, limit poker was essentially solved with a new technique in game theory that allowed them to find a simplified model.

http://spectrum.ieee.org/automaton/robotics/artificial-intel...

Re: AI Beats Four Top Poker Players

#85

Post by davidivadavid »

brilliantcode wrote: post #73
deepnotderp wrote: post #44

Earlier quoted context omitted.

> "So the question is, how does an average joe hacker like me exploit and leverage this wonderful thing called deep learning? I'm not interested in reading PHD papers with advanced calculus." FFS, if you're not willing to read a paper with BASIC CALCULUS (it's HS/college, not advanced like fractional), then I'm not sure machine learning is the right place for you.

Sure it is. Much like most developers don't spend their time reading on computing science papers to do their job, I question why such academic rigour is to be demanded from somebody trying to capitalize on the arbitrage opportunity. Math is good but my time won't be best used if I have to learn calculus all over again just to begin understanding the linguo. Rather have a generic model of what to use and when, hire th…

How do you suggest you take advantage of an arbitrage opportunity if you have 0 informational insight?

Re: AI Beats Four Top Poker Players

#86

Post by gexla »

Do a search for "long term" in online poker and you'll find that the suggestion for players to determine their level of play based on that "long term" is something like one million hands. That's running 4 tables for full days of poker over a long period of time. For sure it's playing more than 4 players.

The probability space of poker is such that 4 competitors isn't going to tell you much.

And defining "top" is difficult because "top" may be more celebrity than anything. Everyone has their different objectives. If you are "top" then you sure as didn't get there by building a case history against AI poker bots. Give these guys a chance to adjust, and give them a chance to figure out why the effort to adjust might be worth bothering with.

Re: AI Beats Four Top Poker Players

#87

Post by brilliantcode »

deepnotderp wrote: post #83
brilliantcode wrote: post #73

Earlier quoted context omitted.

Sure it is. Much like most developers don't spend their time reading on computing science papers to do their job, I question why such academic rigour is to be demanded from somebody trying to capitalize on the arbitrage opportunity. Math is good but my time won't be best used if I have to learn calculus all over again just to begin understanding the linguo. Rather have a generic model of what to use and when, hire th…

> "...academic rigor..." "...calculus.." Dude, I'm not even asking you for anything that difficult here, how would you like it if I walked in and said "well, I don't need to learn any programing, my time is better spent magically making it work".

I found calculus and academia generally very difficult. I'm happy to know that you don't suffer from this terrible affliction which makes studying difficult.

You don't need to know programming if you understand your workers and they understand it enough to convey to you the time & resource constraints.

I have a million dollars for this project. What are the features and functions that will maximize my returns? Hire a couple of smart engineers who already holds the technical knowledge, able to communicate when asked all the pitfalls, potential roadblocks and unknowns so that their productivity is maximized. Set expectations on how they will be able to trade their time for money. At the end of the day, it's a naked call option with limited risk and unlimited reward by utilizing the correct knowledge workers.

Why spend time learning the reference when you can look it up or hire somebody who already has a distilled version in their head which you can utilize?

Re: AI Beats Four Top Poker Players

#89

Post by jedberg »

panorama wrote: post #78
jedberg wrote: post #63

Earlier quoted context omitted.

Anybody here? http://pokerdb.thehendonmob.com/rankings/ Or any of these people: https://en.wikipedia.org/wiki/List_of_World_Series_of_Poker_...

I like your posts, but you're misinformed on this topic. The pros selected are among the best HUNL players in the world right now. Offering the top winners in a separate variant not only misinterprets how different the games are but also falsely equivocates being the best with making the most money. If someone asked you who the best mobile app developers were in the world, your list should comprise solely of mobile a…

You make a fair point. I clearly didn't know what I was talking about. I edited my original post to reflect that.

Re: AI Beats Four Top Poker Players

#90

Post by deepnotderp »

brilliantcode wrote: post #82
ori_b wrote: post #76

Earlier quoted context omitted.

Because this is the language of the domain. Much like code is the language of programming. This isn't academic rigor or lingo. This is how things work. To me, your comment reads similarly to "Much like most business people don't spend their time reading code to do the job, I question why such academic rigor is to be demanded by someone who is translating business requirements into database operations".

I think we are after different things. My goal is not to understand the underlying technology at the academic level but just enough to make money by utilizing the people that do, just like any good capitalist. You don't need to be an oil rig worker to understand where the opportunity lies is what I'm saying. It's possible to understand systems from a blackbox point of view and still be able to exploit them. Afterall,…

If you want to be a venture capitalist, then you need to bring money to the table....
Post reply on HN