Title image source | Xiaoice company
The competition in sports competitions has long ceased to be just the input of the athletes themselves.
Just a few days ago, the Chinese aerial skills sports team, including Xu Mengtao, Jia Zongyang, Qi Guangpu, etc., won a total of "two gold and one silver" in the history of the Beijing Winter Olympics freestyle skiing aerial skills competition, which is also China won the gold medal in the aerial skills project again after winning the first gold medal at the 2006 Turin Winter Olympics, creating a new history with leapfrog development.
The key to this historic breakthrough is that the General Administration of Sport has made a correct decision to focus on special projects related to the Winter Olympics of science and technology, fully trust in the talent and tenacity of athletes, fully trust in the scientific training guidance of coaches, and also believe in the power of artificial intelligence technology.
Among them, the artificial intelligence referee and coaching system "Guanjun" developed by Xiaoice Company quickly attracted people's attention. Because this is the first case of artificial intelligence participating in the training of the world's top competitions and assisting in achieving impressive results.
As an "old project" established by the Winter Olympics in 1994, aerial skills are definitely one of the most difficult and dangerous events of the Winter Olympics: athletes have to slide down from the top of the slope of the track, with the help of the snow jumping platform, rush into the air at a speed of 40 kilometers per hour, and then take off to 15 meters to land, and complete the aerial, aerial movements, landing and other actions in one go.
The special competition content also brings difficulties to daily training: due to the certain danger of aerial skills, athletes have a limited amount of training per day; secondly, the judgment of actions in each training has mainly relied on artificial video video playback in the past, but because the speed of athletes doing actions in the air is still very high, the capture of details is very challenging, and the historical training data recorded in all videos is also very difficult to call and compare.
In order to solve these problems and improve and enhance the performance of athletes as much as possible, the Winter Games Center and the Xiaoice team launched the "Xiaoice-International Athletic Scoring System" (X-IASS), the core task of the project is to try to use AI technology to help improve the performance and competitiveness of athletes in aerial skills.
The Xiaoice team began to train with the team in 2019, continuously collected sports data for more than 3 years, pioneered the "small sample, big task" ice and snow sports analysis model (Xiaoice CV Analysis Model for Winter Sports), and finally created the first artificial intelligence referee and coaching system "Guanjun" that meets the professional requirements of international competitive events.
Unlike the engineers we have seen in the past when they operate computer programs and display data with the team, "Guanjun" is highly automated in actual operation, and can directly referee and score through the key data information of multiple dimensions such as the collected movement trajectory, introduction angle, and high distance. The system is far more refined than ever, and the specific movement changes of a certain body part when the athlete is moving in the air can also be analyzed and recorded, such as the angle of the arm joint of the athlete at a certain moment in the air, thus providing the most detailed training support for the aerial skill athlete.
It is worth mentioning that on the basis of each piece of data analysis, "Guanjun" also established a corresponding exclusive sports file for each team member. Each training of athletes is completely datatized and archived, thus forming a continuous comparable and traceable training record data, helping the coaching team to fully grasp the training effect of athletes, and also allowing athletes to be more targeted in the process of specific actions and physical ability improvement, and finally rapidly improving the training effect of athletes in the simple and efficient training process.
"Guanjun" for the excellent judging ability of aerial skill athletes in the training process, but also let him win the recognition of the International Snow Federation, directly introduced to the Beijing Winter Olympics test match as the only competition referee in the air skills project. In the process of this refereeing, "Guanjun" completed all the refereeing work of a total of 44 people, and the accuracy and fairness of the referee results were unanimously recognized by the International Snow Federation, the Winter Games Center, the team leader and the coaching team, and successfully completed the task of independent refereeing for the first time in the world.
At the beginning of the project, the Xiaoice team understood that this challenge was not small. In the following 3 years, Xiaoice sent a technical product team to train with the whole team, and even the product manager reached the refereeing level of the International Snow Federation, which led to the birth of "Guanjun".
The reason why Xiaoice attaches so much importance to this project is that in addition to the top field of the Winter Olympics, there are also attempts to find a new path for artificial intelligence to land.
"Guanjun" is a good demonstration of the ability of the artificial intelligence system to solve specific problems: when athletes and coaches feel that athletes are integrated with the background at night, and it is too difficult for the human eye to identify the details of the action, "Guanjun" can clearly track the athletes at a speed of 40 kilometers per hour at a distance of 25 meters, and identify and analyze all the action details.
In 2011, in "Penalty Shootouts For Gold," professional baseball manager Billy Bean used data models and data analytics to help players reach their full potential, creating a record of 20 consecutive wins. The reality of 11 years later has been more exciting than the movie, and the artificial intelligence referee and coaching system has successfully assisted the air skills team to win the Winter Olympic championship in more than three years.
Nowadays, artificial intelligence systems such as Guanjun have been increasingly used in the sports industry, such as helping athletes train scientifically, analyzing and predicting events, improving fan interaction, and providing customized peripherals. What is more worth looking forward to is that the future of the sports industry with the help of artificial intelligence, from talent mining, to pre-game preparation, in-game activities, post-game analysis, to improving the media and fan experience, as well as enhancing the management and operational efficiency of sports institutions, the application of artificial intelligence technology in the sports industry has unlimited potential.
End