Multiple Random Masking Autoencoder Ensembles for Robust Multimodal Semi-supervised Learning
This work addresses data reliability issues in Earth observation for environmental monitoring, though it appears incremental as it builds on existing autoencoder and ensemble methods.
The paper tackles the problem of predicting missing data in multimodal Earth observation tasks by proposing an ensemble of autoencoders with random masking, achieving robust performance in semi-supervised learning scenarios.
There is an increasing number of real-world problems in computer vision and machine learning requiring to take into consideration multiple interpretation layers (modalities or views) of the world and learn how they relate to each other. For example, in the case of Earth Observations from satellite data, it is important to be able to predict one observation layer (e.g. vegetation index) from other layers (e.g. water vapor, snow cover, temperature etc), in order to best understand how the Earth System functions and also be able to reliably predict information for one layer when the data is missing (e.g. due to measurement failure or error).