Skip to main content Skip to footer

Alumni Festival at Clare Hall explores the use of AI for football tactics

4 October 2024 College life

We were delighted to host a discussion on the use of artificial intelligence for football tactics as part of the University’s 2024 Alumni Festival, which brings together alumni from around the world to discover the latest fascinating research and reconnect with fellow alumni.
Dr Petar Veličković, Staff Research Scientist at DeepMind and Affiliated Lecturer at the University of Cambridge

On September 28, Clare Hall Associate Dr Petar Veličković gave a talk to a full audience in the Richard Eden Suite on using AI for football tactics. After an introduction from President Alan Short, Dr Veličković shared his research background and projects that he has worked on as a Staff Research Scientist at Google DeepMind. Much of this research has been recognised by international news and publications, such as using AI to improve travel-time predictions in Google Maps, and advancing mathematics by guiding human intuition with AI.

The main portion of the talk was dedicated to the collaboration between DeepMind and Liverpool FC to develop TacticAI, an AI football tactics assistant. TacticAI incorporates both a predictive and a generative component, allowing the coaches to effectively sample and explore alternative player setups for each corner kick routine and to select those with the highest predicted likelihood of success. Dr Veličković and his team validated TacticAI on a number of relevant benchmark tasks: predicting receivers and shot attempts and recommending player position adjustments. The utility of TacticAI is also validated by a qualitative study conducted with football domain experts at Liverpool FC. They show that TacticAI’s model suggestions are not only indistinguishable from real tactics, but also favoured over existing tactics 90% of the time, and that TacticAI offers an effective corner kick retrieval system. TacticAI achieves these results despite the limited availability of gold-standard data, achieving data efficiency through geometric deep learning.

We thank Dr Veličković and attendees who contributed to a wonderful event, and look forward to hosting similar events for alumni in the near future.