Layer-Wise Relevance Propagation for Echo State Networks Applied to Earth System Variability
Marco Landt-Hayen, Peer Kröger, Martin Claus, Willi Rath · 2022
Artificial neural networks (ANNs) are powerful methods for many hard problems (e.g. image classification or time series prediction). However, these models are often difficult to interpret. Layer-wise relevance propagation (LRP) is a widely used technique to understand how ANN models come to their conclusion and to understand what a model has learned. Here, we focus on Echo State Networks (ESNs) as a certain type of recurrent neural networks. ESNs are easy to train and only require a small number of trainable parameters. We show how LRP can be applied to ESNs to open the black-box. We also show an efficient way of how ESNs can be used for image classification: Our ESN model serves as a detector for El Niño Southern Oscillation (ENSO) from sea surface temperature anomalies. ENSO is a well-known problem. Here, we use this problem to demonstrate how LRP can significantly enhance the explainablility of ESNs.