noambrown

9 ms·

31.
by noambrown 7y ago
We give more details on this in the supplementary material.
32.
by noambrown 7y ago
I don't have the exact percentages but I think it's less than 10%. It's not really possible to measure the bot's performance just in specific situations, but my feeling is the bot performs relatively well in these situat
33.
by noambrown 7y ago
There were several improvements but the most important was the depth-limited search. Libratus would always search to the end of the game. But that's not necessarily feasible in a game as complex as six-player poker. With these new algo
34.
by noambrown 7y ago
We tried to make the paper as accessible as possible. A lot of these questions are covered in the supplementary material (along with pseudocode). - Are the action and information abstraction procedures hand-engineered or learned in some ma
35.
by noambrown 7y ago
The number of players is kind of arbitrary given the techniques we're using. We chose 6 because that's the most popular/common format for poker. I don't think there's any scientific value in also doing 10.
36.
by noambrown 7y ago
LLinusLove was one of the players. Chris Ferguson was in one of the 5 AI's + 1 Human experiment but not the 5 Humans + 1 AI experiment. We used AIVAT to reduce variance, which reduces the number of samples we need by roughly a factor o
37.
by noambrown 7y ago
I think it took the community a while to come up with the right algorithms. So much of early AI research was focused on beating humans at chess and later Go. But those techniques don't directly carry over to an imperfect-information ga
38.
by noambrown 7y ago
We're doing cash games in this experiment. At the end of the day, this is about advancing AI, not about making a poker bot. Going from two-player to multi-player has important implications for AI beyond just poker. I don't think t
39.
by noambrown 7y ago
The bot does bluff, and in fact it learns from self-play that bluffing is (sometimes) the optimal thing to do. At the end of the day, bluffing is simply betting when you have a weak hand. The bot learns from experience that when it bets wit
40.
by noambrown 7y ago
There is some open-source code in this area, and hopefully there will be more going forward. Here's one example: https://github.com/EricSteinberger/Deep-CFR
41.
by noambrown 7y ago
1) The goal was to show convincingly that we could handle multi-player poker. The exact number of players was kind of arbitrary. We chose six-player because that's the most common/popular format. Considering training the 6-player
42.
by noambrown 7y ago
In the paper we include a graph of performance over the course of the 10,000-hand 5 humans + 1 AI experiment that was played over 12 days. There's no indication that the bot's performance decreased over time (there is a temporary
43.
by noambrown 7y ago
We considered both options but decided to go with 100BB because that is the standard in the poker world. It doesn't make a big difference for these techniques though.
44.
by noambrown 7y ago
We played 10,000 hands of poker over the course of 12 days in the 5 humans + 1 AI experiment, and 5,000 hands per player in the 1 human + 5 AI's experiment. That's a good amount of time for a player to find a weakness in the syste
45.
by noambrown 7y ago
I don't think the poker world would be happy with us if we did that. Heads-up limit hold'em isn't really played professionally anymore, but six-player no-limit hold'em is very popular.
46.
by noambrown 7y ago
Training was super cheap. It would cost under $150 on cloud computing services. The training aspect has some improvements but is at its core similar to Libratus. The search algorithm is the biggest difference. There aren't that many gr
47.
by noambrown 7y ago
No, it wouldn't be hard. We chose six players because it's the most common/popular form of poker. Also, as you add more players it becomes harder and harder to evaluate because the bot's involved in fewer hands, we need
48.
by noambrown 7y ago
Thanks!
49.
by noambrown 7y ago
We talk about this a bit in the paper. Based on the feedback from the pros, the bot seems to "donk bet" (call and then bet on the next round) much more than human pros do. It also randomizes between multiple bet sizes, including v
50.
by noambrown 7y ago
The bot handles each hand independently. How the players play in one hand does not affect how the bot plays in future hands at all. That said, it did train by playing against itself (before the experiment against the humans began).
51.
by noambrown 7y ago
The paper is here: https://science.sciencemag.org/content/early/2019/07/10/scie... It's going for game-theory-optimal play. It doesn't adapt to its opponents' observed weaknesses. But
52.
by noambrown 7y ago
We're really focused on advancing the fundamental AI aspect. We're not here to kill poker. The popular poker sites have quite sophisticated anti-bot measures, but it's true that this is an arms race.
53.
by noambrown 7y ago
Thanks! I think going beyond two-player/team zero-sum games is really important. This was a first step, but it's definitely not the last. I'm hoping to continue in this direction, and maybe start looking at interactions invol
54.
by noambrown 7y ago
A little bit of both. We didn't think we needed the extra computing power. And we really wanted to convey how cheap it is to make a superstrong poker AI with these latest algorithms.
55.
by noambrown 7y ago
It was online. The players were playing from home on their own schedules. The bot did not look at any tells (timing tells or otherwise). The players knew they were playing a bot and knew which player the bot was.
56.
by noambrown 7y ago
I'm one of the authors of the bot, AMA
57.
by noambrown 9y ago
Author of the paper here. This actually depends on the stack sizes. At 200BB deep, 27o is the worst hand, though 23o is not far behind.