- Titre : A Comparative Study of MLM BERT-Based Models for Trajectory Representation Learning
- Présentateur : Amir Badawi (IGN)
- Résumé : The goal of this work is to provide a framework to explore systematically the influence of several implementation choices and factors on the performance of LLM BERT-based models for trajectory representation learning, the performance being evaluated on the MLM (Masked Language Model) task.
- Sujet de recherche associé : Action 134 - Modelling multi-scale trajectories for soft mobility and outdoor activities