Cluster-based Input Weight Initialization for Echo State Networks
This work addresses a specific bottleneck in ESNs for researchers and practitioners in audio, image, and radar recognition, offering an incremental improvement over random initialization.
The paper tackled the problem of random initialization in Echo State Networks (ESNs) by proposing an unsupervised method using K-Means on training data to initialize input connections, resulting in equivalent or superior performance with significantly fewer reservoir neurons across various datasets.
Echo State Networks (ESNs) are a special type of recurrent neural networks (RNNs), in which the input and recurrent connections are traditionally generated randomly, and only the output weights are trained. Despite the recent success of ESNs in various tasks of audio, image and radar recognition, we postulate that a purely random initialization is not the ideal way of initializing ESNs. The aim of this work is to propose an unsupervised initialization of the input connections using the $K$-Means algorithm on the training data. We show that for a large variety of datasets this initialization performs equivalently or superior than a randomly initialized ESN whilst needing significantly less reservoir neurons. Furthermore, we discuss that this approach provides the opportunity to estimate a suitable size of the reservoir based on prior knowledge about the data.