AI is the most popular team to play football when

AI teams up to play football and become a football coach, and AI invades the football world

recently, deepmind opened source the robot football simulation environment mujoco soccer, which realizes the simulation of 2v2 football match

although the appearance of players is relatively simple (also a ball), deepmind enables them to find teamwork in intensive learning. Friends who love football games seem to smell its prospect: you should go to EA to cooperate with FIFA games

similar to alphago, deepmind also trains many "players". Deepmind selected 10 pairs of football teams, which were made from different training plans

each of these 10 teams has 25 billion frames of learning experience, and deepmind has collected 1million games between them

let's take a look at one of the 2v2 football games from a bird's-eye view:

deepmind found that with the increase of learning, "players" gradually changed from "lone Rangers" to individuals with teamwork spirit

at the beginning, the blue No. 0 player always carries the ball by himself, regardless of the position of his teammates. After 80 billion pictures of training, it has learned to actively look for opportunities for passing cooperation, which will also be affected by the position of its teammates

in one of the games, we can even see two consecutive passes between teammates, that is, the 2-1 pass coordination that often occurs in human football games

in addition to individual skills, deepmind's experimental results have also been defeated by tactics in the football world

among the 10 agents selected in the experiment, B is the strongest, with an ELO score of 1084.27; The second is C and the wear resistance test, with an ELO score of 1068.85; A's score of 1016.48 ranks fifth among them

according to the calculation rules of ELO score, we will mistakenly think that the winning rate of B against a should reach 62%. In fact, a can win or draw B in 59.7% of the games

the above figure shows the video of the game between agents a, B and C, which qualitatively shows the diversity of tactical strategies of most enterprises in football, whose financing range is between 5million yuan and 50million yuan

they will ask the AI coach to arrange the starting lineup of the game and choose the attack and defense tactics of the team in the game against the white hawks on February 9 local time. Will this be the beginning of a new era of football

Wingate finrich became the first football team to use AI coaches

the working mode is similar to that of human beings

AI football coach is a system created by the expert team "GreenShot" in this field. It can formulate corresponding pre game and in game strategies according to different opponents through data analysis and other methods. Its working mode is very similar to that of human coaches

Wingate finridge, the fish belly team that currently plays in England's seventh level team, decided to use an artificial intelligence coach to become the first person to eat crabs

"whether the coach is a virtual AI or a human coach, the principle of football is the same." Team manager Norman made no secret of his love for AI in an interview

"equipped with AI in the team's coaching team, I think it will complement our tactics."

Norman believes that AI coaches can lead the team to success by following some basic principles of the football world

Dixon, the leader of "GreenShot", the developer of artificial intelligence football coaching system, also looks forward to the impact of high technology on British football. After all, in the football world, the application of artificial intelligence coaching is the first time

"we are eager to see what will happen if we introduce an AI football coach into the game. In just a few weeks, our team developed this AI football coach."

"it can easily consult and analyze the battle data of both sides of the game. As time goes by, AI will learn from the collected data and become an insightful tactical aid."

it can adjust according to the progress of the game

so, how does the AI coach work

AI coaches use the intelligent interactive system that is widely used at present, which converts the user's voice into text, and then collates and sends it to the background of the system

on the basis of big data, AI will feed back the most appropriate tactical layout to users through analysis. At the same time, it also has the storage function to learn with the help of big data, so as to give a more reasonable decision in the subsequent competition

the time to use AI coaches is usually before and during the game. Before the start of the game, the team staff will share the physical fitness level, enthusiasm, type and other information of the team's players with AI, and will also provide detailed information about the strengths, weaknesses and key players of the opposing team

then, after analysis, the AI football coach gives the most effective formation and tactics for specific opponents

of course, the ball game changes rapidly, and the result of the game may change in just a few seconds. So in order to help the human coaching team make tactical decisions during the game, AI can also provide suggestions and inspiration for replacing players and game tactics according to the real-time score in the game

for example, if the team falls behind 0-2, the AI coach may suggest replacing the midfield with a striker to adopt a more offensive strategy

sports intelligence has become a trend

does the entry of AI coaches into the football industry mean that the career of football coaches will be "subverted" in the future? The reality may not be so cruel

it is understood that the role of this technology in the Wingate finrich team is only temporary, and its ultimate goal is to allow everyone to download and interact with it. However, both professional football managers and teachers in charge of school teams can be helped by using such "Ai coaches" in the future

the trend of sports becoming more intelligent is difficult to stop. As Microsoft CEO Satya NADELLA said earlier, "I'm afraid there will be no other industry in the world that is so completely subverted by data and electronic technology like sports."

since 2006, companies have begun to monitor and record various data of athletes, hoping to use these data to analyze and decide whether to replace personnel, how to avoid athletes' injuries and organize training

the same is true of NBA, the world's top basketball event. Ken Di Gennaro, the vice president in charge of IT application in the alliance, revealed that Nb should stop operation in time and get in touch with technicians in time. A has established a complete data computing system, which can mine data in competitions that human beings cannot quantify, and conduct data modeling through machine learning

recently, Microsoft has launched a sports performance platform system, which can analyze athletes' performance in training and competition, and provide athletes with sports based analysis data and solutions

at present, the Seattle imperial women's football team, royal society, Benfica and the Australian cricket team have begun to use this system

"for AI technology that forms a vicious circle, the sports industry is a fertile field with many opportunities. Whether it is business operations or personal data analysis of players, it is suitable for the application of AI."