English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Sophisticated trajectory prediction models that effectively mimic teamdynamics have many potential uses for sports coaches, broadcasters andspectators. However, through experiments on soccer data we found that it can besurprisingly challenging to train a deep learning model for player trajectoryprediction which outperforms linear extrapolation on average distance betweenpredicted and true future trajectories. We propose and test a novel method forimproving training by predicting a sparse trajectory and interpolating usingconstant acceleration, which improves performance for several models. Thisinterpolation can also be used on models that aren't trained with sparseoutputs, and we find that this consistently improves performance for all testedmodels. Additionally, we find that the accuracy of predicted trajectories for asubset of players can be improved by conditioning on the full trajectories ofthe other players, and that this is further improved when combined with sparsepredictions. We also propose a novel architecture using graph networks andmulti-head attention (GraN-MA) which achieves better performance than othertested state-of-the-art models on our dataset and is trivially adapted for bothsparse trajectories and full-trajectory conditioned trajectory prediction.

Enhancing Trajectory Prediction using Sparse Outputs: Application to Team Sports