View: https://twitter.com/GoogleDeepMind/status/1770121564085707082
Subscribe to read
Google DeepMind has developed a prototype artificial intelligence football tactician in collaboration with Premier League club Liverpool, in the latest push to use the technology to master the ebb and flow of big-money sports.
The computerised coach’s suggested improvements to players’ positions at corner kicks — a large potential source of goals — mostly won approval from human experts, according to a paper published in Nature Communications on Tuesday.
DeepMind, which has previously used its algorithms to crack difficult board games such as Go, said patterns seen on sports fields could also offer lessons on how to apply AI in other areas such as robotics and traffic co-ordination.
On the pitch, the company’s TacticAI system reflects both the possibilities and current limitations of intensive efforts to use AI to gain a sporting edge beyond that offered by existing data analysis methods.
The technology promises benefits in planning for situations with predictable starting points, such as corners. The wider task is to apply it to the richer variability of open play.
“What’s exciting about it from an AI perspective is that football is a very dynamic game with lots of unobserved factors that influence outcomes,” said Petar Veličković, a DeepMind researcher and co-author of the Nature paper. “It’s a really challenging problem.”
The DeepMind project is the product of three years of work with Liverpool on deploying AI, including in areas such as penalty kicks and predicting movements of players.
DeepMind’s latest model uses geometric deep learning on a data set comprising 7,176 corner kicks from the English Premier League between 2020 and 2023. Corner kicks represent a significant opportunity for attacking teams: along with other so-called set pieces, such as free kicks, they account for about 30 per cent of all goals.
How TacticAI suggests improved defences against corner kicks
The AI assistant's suggestions are based on data from more than 7,000 corners
TacticAI analysed outcomes from corner kicks with various configurations of players, using criteria such as who received the ball and whether they were able to shoot. It then suggested positional improvements and assessed their plausibility and usefulness in a blind case study by five experts at Liverpool: three data scientists, one video analyst and one coaching assistant.
The experts could not distinguish the AI-generated scenarios from actual match situations, the researchers said, favouring the TacticAI advice 90 per cent of the time over existing strategies. This showed the tool “readily provides useful, realistic and accurate” suggestions, the paper said.
Liverpool did not respond to a request for comment on whether it had implemented any of TacticAI’s suggested changes, as manager Jürgen Klopp strives to end his tenure on a high note with trophies this May.
The use of data analysis to improve outcomes in sport has grown ever more sophisticated since Michael Lewis brought it to wide attention with his 2003 book Moneyball: The Art of Winning an Unfair Game. Lewis recounted how the Oakland Athletics baseball team used novel measures of player attributes to compete against better-funded rivals.
AI’s growing capabilities have now stoked interest in its potential uses across sport. The US National Football League and Amazon Web Services have developed a player health tool known as the Digital Athlete, which they hope may in time be able to predict and prevent injuries.
Projects such as TacticAI give clues as to how AI in football might evolve, said Sudarshan Gopaladesikan, director of football intelligence at Italian top division team Atalanta.
“This is the way that AI can help us approach football in a chunked or categorical way — as opposed to thinking it’s just this one big continuous flow and we don’t know what’s going on,” he said.
