Deep Representation Learning for Personalised High Granularity Cycling Performance Prediction (2021)

Abstract

Modelling the behaviour and performance of cyclists during competition ahead of time has several useful applications to professional cycling teams. Although existing physics-based approaches perform reasonably well in this regard, they generally require several assumed parameters. Performance is also dependant on several variables which may be hard to quantify, such as a rider’s cornering ability and pacing strategy. The work presented in this thesis is a fully data-driven approach to speed pre- diction in time trials performed by elite cyclists. Inspired by FaceNet [1], siamese neural networks are trained to transform previous efforts of cyclists into a low dimensionality embedding space, clustering embeddings on the rider that per- formed them. Embeddings are used alongside other contextual information as features in a recurrent neural network architecture that is able to predict a rider’s speed along a formerly unseen course. The addition of embeddings allow models to be trained on data from multiple riders and leverage more features without overfitting, resulting in a significant overall increase in predictive performance compared to a similar baseline architecture trained on individual riders.