A General Framework for Hypercomplex-valued Extreme Learning Machines

TL;DR

This paper introduces a comprehensive framework for hypercomplex-valued extreme learning machines (ELMs), enabling efficient processing of high-dimensional data across various hypercomplex algebras with demonstrated superior performance.

Contribution

It develops a general framework for hypercomplex-valued ELMs, including derivation of learning algorithms and experimental validation on time-series and image data.

Findings

Hypercomplex-valued ELMs outperform real-valued models in high-dimensional tasks.

The framework supports various hypercomplex algebras, including four-dimensional examples.

Experimental results show excellent performance in time-series prediction and image auto-encoding.

Abstract

This paper aims to establish a framework for extreme learning machines (ELMs) on general hypercomplex algebras. Hypercomplex neural networks are machine learning models that feature higher-dimension numbers as parameters, inputs, and outputs. Firstly, we review broad hypercomplex algebras and show a framework to operate in these algebras through real-valued linear algebra operations in a robust manner. We proceed to explore a handful of well-known four-dimensional examples. Then, we propose the hypercomplex-valued ELMs and derive their learning using a hypercomplex-valued least-squares problem. Finally, we compare real and hypercomplex-valued ELM models' performance in an experiment on time-series prediction and another on color image auto-encoding. The computational experiments highlight the excellent performance of hypercomplex-valued ELMs to treat high-dimensional data, including models based on unusual hypercomplex algebras.