[ad_1]
In March 1950, Charles Reep, a Royal Air Force captain and well-trained accountant, turned to football. Reep became interested in the sport in the 1930s and was fascinated by Herbert Chapman’s pioneer Arsenal team. He returned from World War II and found The tactical revolution he had witnessed had stalled.
In the end, in the three-division half of the singles between Swindon Town and Bristol City, Repp ran out of patience. During this period, he watched countless attacks pointlessly. He grabbed a notebook and a pencil, and started to jot down everything that happened on the court: he started counting the number of passes and shots, which was the first systematic attempt to analyze football using data.
Seventy years later, the data revolution has reached the grassroots level-Fan fluent G And the net expenditure, the senior team directly selects PhD students in statistics from the university to find the advantage. Now, Liverpool, defending the Premier League championship, has joined forces with DeepMind to explore the use of artificial intelligence in the football world.Papers published today by researchers from these two organizations Journal of Artificial Intelligence ResearchSome potential applications are outlined.
“The timing is right,” said Karl Tuyls, an AI researcher at DeepMind and one of the paper’s lead authors. DeepMind’s cooperation in Liverpool stems from his previous experience at the University of Liverpool. (Founder of DeepMind Demis Hassabis He was still Liverpool’s life and an advisor for this study. ) The two groups got together to discuss how AI can help football players and coaches. Liverpool also provided DeepMind with data on every Premier League game the club played from 2017 to 2019.
In recent years, the amount of data available in football has exploded by using sensors, GPS trackers, and computer vision algorithms to track the movement of players and balls. For football teams, artificial intelligence provides a way to spot patterns that coaches cannot do. For DeepMind researchers, football provides them with a limited but extremely challenging environment for them to drive test the algorithm. “Like [soccer] It’s super interesting because there are many agents and there are competition and collaboration issues,” Tuyls said. Unlike chess or Go, football has inherent uncertainty because it is played in the real world.
However, this does not mean that you cannot make predictions, and this is where AI may be particularly useful. This article demonstrates how to train a model on data about a specific team and lineup to predict how its players will react in a specific situation: If you hit a long ball on the right-hand channel against Manchester City, for example, Kyle Walker Will move in a certain direction, and John Stones may do other things.
This is called “ghosting” because the alternative trajectory is overlaid on what is actually happening (for example in a video game) and has a range of different applications. For example, it can be used to predict the meaning of tactical changes or how opponents will play after a key player is injured. These are things that coaches might notice, and Tuyls emphasized that the goal is not to design tools to replace them. He said: “There is a lot of data, and there is a lot of data that needs to be digested. Processing these massive amounts of data may not be so easy.” “We are trying to establish assistive technologies.”
As part of the paper, the researchers also analyzed the 12,000 penalty kicks performed in Europe in the past few seasons-according to the players’ style of play, the players are divided into several categories, and then based on this information, they can make predictions that they may be punished. And whether they are likely to score. For example, strikers are more likely to aim at the lower left corner than midfielders. The latter uses a more balanced approach. Data shows that for free throwers, the best strategy is to use their strongest strength. This is not surprising. .
[ad_2]
Source link
