Slashdot Mirror
OpenAI Is Beating Humans At 'Dota 2' Because It's Basically Cheating (vice.com)
Motherboard's Matthew Gault provides another possibility for how OpenAI's bots managed to beat professional human players in two consecutive games of Data 2. Gault argues that "it was only possible thanks to significant guardrails and an inhuman advantage" -- not necessarily because the AI was more clever than the humans. From the report: The OpenAI Five bots consisted of algorithms known as neural networks, which loosely mimic the brain and "learn" to complete tasks after a process of training and feedback. The research company put its Dota 2-playing AI through 180 days worth of virtual training to prepare it for the match, and it showed. However, the bots had to play within some highly specific limitations. Dota 2 is a complicated game with more than 100 heroes. Some of them use quirky and game-changing abilities. For this exhibition, the hero pool was limited to just 18. That's an incredible handicap because so much of Dota 2 involves a team picking the proper group composition and reacting to what its opponents pick. Reducing the number of champions from more than 100 to 18 made things much simpler for the AI.
The OpenAI Five bots also played Dota 2 by reading the game's information directly from its application programming interface (API), which allows other programs to easily interface with Dota 2. This gives the AI instant knowledge about the game, whereas human players have to visually interpret a screen. If a human was able to do this in a competitive match against other humans, we'd probably call it cheating. Even with this AI advantage, Walsh and his team beat the bots in the third game, when the match organizers turned hero selection over to the crowd, which gave the AI a weak hero composition. Walsh thinks he and his team could eventually beat the AI in a fair right, even given the limited hero pool and other restrictions.
The OpenAI Five bots also played Dota 2 by reading the game's information directly from its application programming interface (API), which allows other programs to easily interface with Dota 2. This gives the AI instant knowledge about the game, whereas human players have to visually interpret a screen. If a human was able to do this in a competitive match against other humans, we'd probably call it cheating. Even with this AI advantage, Walsh and his team beat the bots in the third game, when the match organizers turned hero selection over to the crowd, which gave the AI a weak hero composition. Walsh thinks he and his team could eventually beat the AI in a fair right, even given the limited hero pool and other restrictions.
In other important news for nerds, the horse runs faster than a steam engine...
what many of us pointed out in the comments when the story originally ran here on /.
I find applications of machine assistance more promising. I'm still waiting for a way to program the original starcraft so that the AI can manage the tedious resource management while a human player can work on strategy. Even something like self driving cars make more sense as an assist. For someone who is elderly do they really care if the car is being driven by a computer or by a human sitting in an office somewhere. If the human sitting in the office gets the advantage of an AI highlighting the road for them and the elderly only has to pay the human for the 10 minutes to and from the appt. The reason chauffeurs are so expensive is because you have to pay them to sit. If you have AI drive the interstates and a human drive the final 10 minutes on each end, it would be very economical.
Did it submit input directly? Humans can't do that! It should use robotic arms!
Cheating means that true Artificial Intelligence that mimics humans is here!
The shepherds did so well protecting the flock that the sheep no longer believed that wolves existed.
If I were a competitive player I'd see it as an opportunity; if I won I could get mass bragging rights and if not then I'd learn try different strategies, just like the AI would.
This is another shit tier "article" from vice. For fucks sake, can they ever hire/pay decent people to write articles?
The neural networks being trained here do nothing revolutionary. The only thing they do is take multiple inputs and optimise for outputs that give a higher rate of wins over a number of games played. Nothing more, nothing less. Humans will be inferior at such an endeavour because they simply can't account for so many input values.
No, humans could never beat a sufficiently trained NN at anything, because they simply can't. Alpha Go proved it. Alpha chess zero proved it. Leela chess zero proved it. Given enough computing power, a self learning NN will beat humans at any task.
Dota is a complicated game but its a 5v5 and a hive mind will always be more organized than five individuals.
More importantly there are items to Hex, Silence, Disarm and otherwise disable other characters and a fight can be decided on who gets disabled first.
Human beings have mental reaction time and physical processing time of moving a mouse to the right coordinates.
Computers have none of that and they will always click on you before you can click on them. That doesn't make them smart.
Really? The best they could come up with is comparing the bot to a programmable mouse, illegal in human play? This standard basically disqualifies any bot implementation being fair in their definition just on the basis of being a program. What do they want? The bot to send commands to a human operator which plays the game for it?
Their slightly less ridiculous claim is saying the bot is unfair because it interfaces with the game not through mouse and keyboard and the dota renderer but through valve's bot api. Just..fine, if you dont care about how good a bot can be under the dota developer's own definition, then ignore the results.
I just dont understand the mindset here, is this resistance to change? People dont want AI to be good, so they make up narratives to explain away recent successes? It really feels like the same mindset as conspiracy theorists.
...routinely, for probably close to 3 years now. Still don't have a fucking clue what its actual name is.
"The research company put its Dota 2-playing AI through 180 days worth of virtual training to prepare it for the match"
it played 180 years worth of dota 2 games every single day in preparation for this, not 180 days
"It has the knowledge of where everyone is..."
not true at all
another common thing i hear mentioned is an unfair reaction time, yet they more than doubled it to around professional levels without consequence. it's ability is in coordination
"Reducing the number of champions from more than 100 to 18 made things much simpler for the AI."
that's the goal, making it simpler for the ai. originally there were only 5 heroes, both teams playing the same, no warding, much less items, a single courier, among other things. heroes and items that make illusions or copies or themselves, or control multiple units, these specifically were avoided because it could perfectly macro control them. that alone removes about half the things from dota. they also used an older version of the game. doing the entire base game would take too much time to learn anything or improve. the rate that machine learning is improving means it would be feasible eventually, but trying to do so now would be like building a bridge using only chopsticks when you know in a month you will have actual lumber to work with
the goal for this isn't to crush humans, that's just (inconveniently for players) the only way they have of knowing/confirming its ability. they are using dota 2 as a model for learning complex tasks suitable for real world purposes. they already published the algorithms and training code for this dota experiment, even using the same implementation for something very different like dextrous robot hand manipulation
Do you know the difference between DOTA and DATA?
Neither does BeauHD.
No mention of the fact the courier visibility to the enemy team had to be turned off as well because the AI was so bad at using them?
If you'd been paying any attention at all, you'd know that the AI's latency was set to 200ms, which is larger than the average human's.
Same as these guys -- their own logic is self-contradictory. Either DOTA is a game mainly about reaction time, in which case the 18-player limit will have almost no effect; or DOTA is a game mainly about strategy and how to use characters together, in which case the direct interface will have little effect. Given the fact that poorly-chosen characters caused the computer to lose decisively, I think the first one is much more likely.
The OpenAI team have stated over and over why they use the direct interface rather than scanning the pixels: Because they know how to get an AI to scan pixels, but they don't know how to get an AI to do strategy. So they're focusing their training time on strategy rather than wasting GPU time scanning pixels.
TCP: Why the Internet is full of SYN.
for instance, the performance of AI on simple video games is due, in part, to the fact that it simply has less latency in scanning the screen, pushing that information into a meat cortex, and pushing its meat appendage against an ergonomic device. this doesn't detract, imho, from the accomplishment (however much you may or may not think of it as significant in the first place).
as for cheating: well, i think people took a tech demo maybe a bit too seriously... it's still impressive imho, even if it was playing a "special edition" of dota2. i've seen much, much more incomplete proofs-of-concept get obscene funding.
"They were pure niggers." – Noam Chomsky
Who is Matthew Gault?
Corruption is convincing someone that the selfless ideal is the same as their selfish ideal.
Developers of open AI have clearly acknowledged these limitations in their blog posts. And they have plans to address them in future.
Right now it's a big (PR) win because there is another camp in ML research that is publishing papers about how "RL (reinforcement learning) doesn't work". And here is a RL based systems which achieves never before seen results in real world.
It's only a matter of time before they improve upon it to make it human level. There is no reason why they should just start with the most difficult engineering problem.
It's like complaining that wright brothers plain didn't have a proper toilet!
the game where they were given an awful team was far more interesting to watch and try and understand the AI behaviour. They were forced to try and implement strategies with little synergy and without much hope of victory. Humans would have GG out but the bots, even with 0% predicted chance of winning, were still trying to work out ways to win.
I watched all the matches. The AI 'reading' the API didn't stop them making fucking 'dumb' decisions no human player would ever make. Buying salves(consumable healing item) constantly, wasting smokes randomly, double ward glitching. The only questionable play they made was the near instant hex on a blinking earthshaker about to land his combo. But top level players know what the fuck earthshaker does.
They are prepared for him to show up like that. Its called fucking learning, and its exactly what this AI did. You can talk about unfair reaction speed, but experience means far more, holding your disable cooldown for the right target.
Making the correct decision is not fucking cheating. They acted without hesitation, and they crushed m-dog, f-dog and every other dog that dared taunt them with 'baited kid'. Up until they let 4chans IRC aka twitch chat draft their 3rd game.
Side musing: How much of human "mental processing" involves the reuse of vision-circuits, meaning there's no clear boundary?
I did a spit take when I read "Dota 2 is a complicated game". Starcraft is a complex game. Sure there are 100 heroes in Dota, but only 10 of them can be on the field at once (I think, I don't play it so please correct me if I'm wrong). In Starcraft you can have more than 400 units on the field at once, all doing different directed actions.
I gather from your post you didn't watch the event nor read about the AI.
The AI's didn't have a Hive Mind. The bots had to communicate to each other and had comm-lag tuned to human-lag. They also had reaction-lag tuned to human-lag. For instance, there were several times during the matches two bots would drop wards near each other at almost the same time, just like we humans do.
I agree that they they are able to "click on" the right thing unerringly, but let's swap some words around this statement: They [AIs] will always click on you before you can click on them. That doesn't make them smart. to Pros will always click on you before a Wood Tier player can click on them. It doesn't make them smart. The smartness of the AI and that it trounced the shit out of the pros at this event had less to do with clicking and more to do with its meta play. (It also did some questionable things that normal pros wouldn't do.)
To wrap this up, it isn't done. OpenAI isn't calling victory until it is playing a full match without restrictions. In the event, they discussed why certain decisions were made. For instance, no summon classes because that would be unfair to the humans because OpenAI would micro them better. Carriers couldn't be killed because, just like in product development, you don't always have the time to do everything before product launch (i.e., this event).