Monday, 14 August 2017

AI Crushed a Human at Dota 2πŸ€–

Elon Musk founded OpenAI as a nonprofit venture to prevent AI from destroying the world, something Musk has been beating the drum about for years. Just last month he told a group of US governors that AI represents a “fundamental risk to the existence of civilization.”

The idea that an AI can be trained to beat the best human players at virtually any game shouldn’t be a shock in 2017, but somehow, we’re still not quite used to the idea.  This year’s “The International Dota 2 Championships took place August 7-12 at Key Arena in Seattle, and one of the big events pitted Danylo “Dendi” Ishutin against an OpenAI bot.  The OpenAI bot also defeated other top pros during the course of the event.

Dota 1v1 is a complex game with hidden information. Agents must learn to plan, attack, trick, and deceive their opponents. The correlation between player skill and actions-per-minute is not strong, and in fact, AI’s actions per minute are comparable to that of an average human player.

In this style of play, OpenAI’s bot is completely dominant. Early into the first game, the event’s host asked Dendi: “Does it feel like a player, like a person?” “Uhhhh… nope,” Dendi answered, distracted by yet another perfectly-timed attack. “This guy is really scary,” Dendi kept saying, and at one point yelped “Please stop bullying me!” when the bot rushed at him aggressively.

Bot is trained entirely through self-play,it starts off completely random with no knowledge of the world and simply plays against a copy of itself, which means it always has an evenly matched opponent, and it climbs this ladder of skill level until it's able to reach the performance of the best professional players in the world.

While gaming and esports might not seem like the most obvious avenue for academic development, OpenAI’s bot is the latest example of modern AI applications. With machine learning, bots like OpenAI’s can generally be put into a scenario and independently learn from its past mistakes to become smarter.

From the very beginning, it just keeps playing against a copy of itself. It starts from complete randomness and then it makes very small improvements, and eventually it's just pro level.

OpenAI robo-brain clearly came out of the exhibition looking like the stronger player, but the tech company isn't quite finished. An accompanying video notes that the bot is still a work-in-progress. The goal is to eventually assemble a full team of AI bots for 5v5 matches and, further down the road, to mix AI players in with human players on a single team.

 it is time to put forward a set of rules regulating the very nature of AI.

By the time we are reactive, it’s too late. Normally the way regulation works out is that a whole bunch of bad things happen, there’s public outcry and after many years a regulatory agency is setup to regulate the industry.

AI is a rapidly-developing technology, but it is still far away from self-evolving, almighty software. Facebook uses AI for targeted advertising, whereas Microsoft and Apple use it to power their digital assistants, Cortana and Siri. Google search engine has also been dependent on AI since its inception.


It is time to put forward a set of rules regulating the very nature of AI.


Popular Posts