Tuesday, 18 December, 2018

Dota 2 Pros Play Bots, Get Wrecked

Watch OpenAI's Dota 2 bots square off against five top players live on Twitch today A team of former Dota 2 pros just got owned by AI bots
Cecil Davis | 08 August, 2018, 10:23

After OpenAI took down members of the audience easily during a warmup round, the bot network started and finished the first of the two matches strongly, stopping its human opponents from destroying any of its defensive towers.

Next, OpenAI's developers plan to take the bots to Valve's official worldwide tournament, where they'll compete against the best Dota 2 players in the world.

This is in stark contrast to the simple bots developed by Valve for training and practice purposes, which are typically no match for human players.

Since OpenAI's Dota 2 bot beat professional player Danylo "Dendi" Ishutin at the global 2017, the AI has evolved significantly.

SpaceX relaunches the final Falcon 9 rocket design
The satellite will provide Internet, telephone and cellular backhaul services, according to SSL. That flight is scheduled for launch next month from Vandenberg Air Force Base in California.

The bot technology of OpenAI has reached a point where it can now destroy a team of highly skilled, former professional, Dota 2 players - in a match that wasn't even close.

In spite of the restrictions, though, OpenAI still completely dominated the first two games, through a combination of near-flawless coordination and a "play as five"-oriented playstyle". This resulted in OpenAI Five predicting a lower chance of winning of the match - just 2.9 percent, in fact.

OpenAI's initiatives with Dota 2 began in March a year ago. "And finally, in a game three played purely for pride, the humans managed to squeeze out a win". It then faced a team of some of the world's top players (Blitz, Cap, Fogged, Merlini and Moonmeander) in a best-of-three contest.

After losing to OpenAI Five, Capitalist Walsh expressed his despair in a tweet. OpenAI co-founder and CTO Greg Brockman has previously described it as - if you assume that it takes a human at least 12,000 hours to become a professional at the game, OpenAI's systems essentially play 100 human lifetimes of experience every day to train. Last month, the company released footage of a robotic arm manipulating a cube with uncanny dexterity that was trained using reinforcement learning algorithm as the OpenAI Five. The list of handicaps included the ability to pick only 18 of the 115 available heroes in the pool, while Team Human was given freedom to use any hero of their choosing.