THANK YOU FOR SUBSCRIBING
Bots in the gaming scene have created a lasting impact in a short duration of time. They have not only won against human opponents, but they have also understood and predicted how the human mind makes decisions.
Fremont, CA: Researchers at MIT have designed a bot that is equipped with AI and can defeat a human player in the trickiest way possible. Several gaming bots have been built to compete against human players.
Sometime earlier, a team from Carnegie Mellon University has built the world’s first bot that is equipped to beat professionals in a multiplayer poker game. AlphaGo by DeepMind made it to the headlines in 2016 to bring the best among professional Go players. At present, there are ample of examples where the bots have outperformed professional players in chess and have even combined forces in cooperative online games such as capture the flag. In games like these, the bot has a clear picture as to who its opponent is from the start.
Stay ahead of the industry with exclusive feature stories on the top companies, expert insights and the latest news delivered straight to your inbox. Subscribe today.
In an upcoming conference on Neural Information Processing Systems, the researchers have prepared the presentation of DeepRole, the first gaming robot that can win the online multiplayer game, against the participants, whose team allegiances are unclear and unavailable.
To achieve the win against an unknown player, the bot is programmed with a novel addition to its algorithms called deductive reasoning, which is commonly used for playing poker. The new addition helps the bot to reason the partially recognizable actions and determine the most probable move of the teammate or the opponent. By doing so, it quickly learns who to support and which moves to take for ensuring the victory.
The researchers have put DeepRole through more than 4000 rounds of the online game “The resistance: Avalon” against human players. In the same game, players try to defeat their opponents with secret roles in the progression of the game, meanwhile making sure their role is hidden. DeepRole performed better than the human players consistently as both a teammate and an opponent.
“If you replace a human teammate with a bot, you can expect a higher win rate for your team. Bots are better partners,” states the first author, Jack Serrino ’18, who majored in the fields of electrical engineering and computer science at MIT and has been an avid online “Avalon” player.
These Bots are a part of the researcher of a broader project dealing with the modeling of how humans think and make socially informed decisions. This research can not only help build better robots, but they can also understand, learn, and work along with humans.
“Humans learn from and cooperate with others, and that enables us to achieve together things that none of us can achieve alone,” says co-author Max Kleiman-Weiner, a postdoc in the Center for Brains, Minds and Machines and the Department of Brain and Cognitive Sciences at MIT, and at Harvard University. “Games like ‘Avalon’ better mimic the dynamic social settings humans experience in everyday life. You have to figure out who’s on your team and will work with you, whether it’s your first day of kindergarten or another day in your office.”
More in News