TY - JOUR

T1 - Interactive sequential generative models for team sports

AU - Fassmeyer, Dennis

AU - Cordes, Moritz

AU - Brefeld, Ulf

PY - 2025/2

Y1 - 2025/2

N2 - Understanding spatiotemporal coordination of players in team sports is key to movement models, pattern detection, and computational tactics. Existing generative models propose to capture all stochasticity by a single latent variable and may suffer from entangled representations, or aim to uncover interaction structures of players but then do not focus on their generative ability. As a remedy, we propose a hierarchical latent variable model for predicting trajectories of multiple players. In the generative model, both, discrete role assignments and a latent interaction graph are sampled to allow for different models in subsequent node updates and message passing operations between nodes, where standard Gaussian latent variables are employed per agent and timestep. We cast our approach as a variational autoencoder that provides a disentangled latent space to capture variability in team sport movements and propose a neural architecture for its optimization. We empirically evaluate our approach on tracking data from basketball and soccer and observe that our contribution outperforms the state-of-art in all experiments.

AB - Understanding spatiotemporal coordination of players in team sports is key to movement models, pattern detection, and computational tactics. Existing generative models propose to capture all stochasticity by a single latent variable and may suffer from entangled representations, or aim to uncover interaction structures of players but then do not focus on their generative ability. As a remedy, we propose a hierarchical latent variable model for predicting trajectories of multiple players. In the generative model, both, discrete role assignments and a latent interaction graph are sampled to allow for different models in subsequent node updates and message passing operations between nodes, where standard Gaussian latent variables are employed per agent and timestep. We cast our approach as a variational autoencoder that provides a disentangled latent space to capture variability in team sport movements and propose a neural architecture for its optimization. We empirically evaluate our approach on tracking data from basketball and soccer and observe that our contribution outperforms the state-of-art in all experiments.

KW - Generative models

KW - Graph recurrent neural networks

KW - Soccer

KW - Socio-temporal dependencies

KW - Trajectory forecasting

KW - Informatics

UR - https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=leuphana_woslite&SrcAuth=WosAPI&KeyUT=WOS:001408563700001&DestLinkType=FullRecord&DestApp=WOS_CPL

U2 - 10.1007/s10994-024-06648-2

DO - 10.1007/s10994-024-06648-2

M3 - Journal articles

VL - 114

JO - Machine Learning

JF - Machine Learning

SN - 0885-6125

IS - 2

M1 - 38

ER -