Masked Autoencoder Pretraining for Event Classification in Elite Soccer

Research output: Contributions to collected editions/worksArticle in conference proceedingsResearchpeer-review


We show that pretraining transformer models improves the performance on supervised classification of tracking data from elite soccer. Specifically, we propose a novel self-supervised masked autoencoder for multiagent trajectories. In contrast to related work, our approach is significantly simpler, has no necessity for handcrafted features and inherently allows for permutation invariance in downstream tasks.

Original languageEnglish
Title of host publicationMachine Learning and Data Mining for Sports Analytics : 10th International Workshop, MLSA 2023, Revised Selected Papers
EditorsUlf Brefeld, Jesse Davis, Jan Van Haaren, Albrecht Zimmermann
Number of pages12
Place of PublicationCham
PublisherSpringer Nature Switzerland AG
Publication date26.02.2024
ISBN (Print)978-3-031-53832-2
ISBN (Electronic)978-3-031-53833-9
Publication statusPublished - 26.02.2024
Event10th International Workshop on Machine Learning and Data Mining for Sports Analytics, MLSA 2023 - Turin, Italy
Duration: 18.09.202318.09.2023

Bibliographical note

Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

    Research areas

  • Factorized transformer architecture, Masked autoencoder, Multiagent trajectories, Self-supervised learning, Soccer, Tracking data
  • Informatics
  • Business informatics