Description
A method that randomly mask out all features coming from a specific time-step in time-series data. If the model used is independent of uneven sequences or missing data in time-series, like attention-based transformers, the masked time-steps can be just ignored in the forward prediction of the model. Otherwise, they have to be masked out with some numerical value.
Papers Using This Method
Increasing the Robustness of Model Predictions to Missing Sensors in Earth Observation2024-07-22Lightweight, Pre-trained Transformers for Remote Sensing Timeseries2023-04-27Multi-Modal Temporal Attention Models for Crop Mapping from Satellite Time Series2021-12-14Fine-tuning Handwriting Recognition systems with Temporal Dropout2021-01-31Crop Classification under Varying Cloud Cover with Neural Ordinary Differential Equations2020-12-04