Complex-valued Neural Networks with Non-parametric Activation Functions

TL;DR

This paper introduces a novel fully complex, non-parametric activation function for complex-valued neural networks, leveraging kernel methods to enhance adaptability and performance in tasks like prediction and channel equalization.

Contribution

It presents the first fully complex, non-parametric activation function for CVNNs based on kernel expansions, improving flexibility over traditional split or simple phase-amplitude functions.

Findings

Outperforms real-valued neural networks in key tasks

Demonstrates improved adaptability with kernel-based activation functions

Validates effectiveness through experiments on prediction and channel equalization

Abstract

Complex-valued neural networks (CVNNs) are a powerful modeling tool for domains where data can be naturally interpreted in terms of complex numbers. However, several analytical properties of the complex domain (e.g., holomorphicity) make the design of CVNNs a more challenging task than their real counterpart. In this paper, we consider the problem of flexible activation functions (AFs) in the complex domain, i.e., AFs endowed with sufficient degrees of freedom to adapt their shape given the training data. While this problem has received considerable attention in the real case, a very limited literature exists for CVNNs, where most activation functions are generally developed in a split fashion (i.e., by considering the real and imaginary parts of the activation separately) or with simple phase-amplitude techniques. Leveraging over the recently proposed kernel activation functions (KAFs), and related advances in the design of complex-valued kernels, we propose the first fully complex, non-parametric activation function for CVNNs, which is based on a kernel expansion with a fixed dictionary that can be implemented efficiently on vectorized hardware. Several experiments on common use cases, including prediction and channel equalization, validate our proposal when compared to real-valued neural networks and CVNNs with fixed activation functions.