Online and adaptive pseudoinverse solutions for ELM weights

By A. van Schaik, Jonathan C. Tapson. 2014

51 citations recorded by the source corpus.

The ELM method has become widely used for classification and regressions problems as a result of its accuracy, simplicity and ease of use. The solution of the hidden layer weights by means of a matrix pseudoinverse operation is a significant contributor to the utility of the method; however, the conventional calculation of the pseudoinverse by means of a singular value decomposition (SVD) is not always practical for large data sets or for online updates to the solution. In this paper we discuss incremental methods for solving the pseudoinverse which are suitable for ELM. We show that careful choice of methods allows us to optimize for accuracy, ease of computation, or adaptability of the solution.

Subfields: Artificial Intelligence, Mechanics of Materials

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Mechanics of Materials

Citations in the corpus

Incremental learning algorithms and applications (2016, background): “Due to their simple training, incremental variants can easily be formulated, whereby their reservoir naturally represents rich potential concepts [178, 31, 159, 61].”
Skimming Digits: Neuromorphic Classification of Spike-Encoded Images (2016, methodology): “The linear output weights are iteratively calculated using the Online Pseudo-inverse Update Method (OPIUM; van Schaik and Tapson, 2015), which allows for the iterative calculation of an analytical solution for the weights.”
EMNIST: Extending MNIST to handwritten letters (2017, methodology): “The pseudo-inverse required for the ELM cannot be calculated in a single step due to size of the dataset and the networks are instead trained using the Online Pseudo-Inverse Update Method (OPIUM) [6].”