Google’s AI tool has captured the strategy of football?

Google’s AI tool has captured the strategy of football?

Google’s top AI division, DeepMind, has developed a tool that can predict the outcome of various corner kicks alongside football experts’ tactics, the search giant claims.

Google's AI tool has captured the strategy of football?
Google’s AI tool has captured the strategy of football?

DeepMind has also used artificial intelligence technology to beat human champions of the complex board game Go.

Besides, Google’s technology has also shown a precedent of losing professional gamers in the video game ‘Starcraft 2’.

Google’s London-based AI division showed the first glimpse of an artificial intelligence system called ‘TacticAI’ on Tuesday. They have previously worked with Liverpool Football Club staff to develop this technology for three years.

The new tool combines two types of AI technology. One is ‘generative AI’ and the other is ‘predictive AI’, which claims to achieve ‘state-of-the-art results’, a DeepMind blog post explains about the technology.

According to a study detailing the technology, after conducting a ‘blind test’ of the strategies generated by the AI system, they were found to be indistinguishable from human-designed strategies.

Despite such success, DeepMind says the tool is not designed to take the place of human coaches, but rather to act as an assistant to them.

With the help of TacticAI’s generative model, coaches will benefit from re-organizing corner kick strategies, while increasing the probability of accurate results in some cases. For example, taking a shot in a defensive setup reduces its probability, says DeepMind.

And from this facility, coaches can also identify various important patterns of play.

It can also identify the success or failure of key players in various tactics faster than ever before.

The system has been trained on data from 7,176 corner kicks from the English Premier League in the 2020/21 season, allowing coaches to simulate and analyze the setups of different players during set pieces.

In addition, DeepMind says, by developing this technology, strategic advice can be provided in other games as well.

It can also be used in areas such as robotics and traffic control by translating general information about human psychological states.