Machine learning systems have been struggling with humanity for more than 10 years. Although in its infancy, the results of the tests are truly fascinating. For testing purposes, the 2 species, which shared their trump cards on the games, came face to face this time in the game called Diplomacy.
First released in 1959, Diplomacy is a more refined version of the board game RISK, where 2 to 7 players take on the role of the European power and their opponents try to win the game by conquering their territory. Unlike RISK, Diplomacy requires players to negotiate with each other – forming alliances or betraying each other – before moving their pieces at the same time in the next game phase. The skills of reading and manipulating opponents, persuading players to form alliances and plan complex strategies, managing delicate partnerships, and knowing when to switch sides are all a big part of the game, while these are skills that machine learning systems so far have often lacked.
Last week, Meta AI researchers announced that they’ve overcome these machine learning shortcomings with CICERO, the first AI to perform at a human level in Diplomacy. The team trained Cicero on 2.7 billion parameters for 50,000 rounds on webDiplomacy.net, an online version of the game, and doubled the average score of their opponents in a 5-game league tournament, placing 2nd (out of 19 participants).
The AI ”has proven so adept at using natural language to negotiate with humans in Diplomacy that often other players have chosen to work with CICERO over human participants,” the Meta team said in a press release.
It seems that in the near future we will witness that artificial intelligence will take the lead in the game. Well, do you think this will be the only victory of artificial intelligence developed with machine learning against human beings?
21674
