Universal approximation with complex-valued deep narrow neural networks
This work addresses the theoretical foundations for complex-valued neural networks, which is incremental but important for applications in signal processing and physics.
The paper tackles the problem of determining which activation functions enable complex-valued deep narrow neural networks to approximate continuous functions universally, showing they are universal if and only if the activation is neither holomorphic, antiholomorphic, nor ℝ-affine, with sufficient widths such as 2n+2m+5 and n+m+3 for subclasses, and provides quantitative bounds for smooth non-polyharmonic activations.
We study the universality of complex-valued neural networks with bounded widths and arbitrary depths. Under mild assumptions, we give a full description of those activation functions $\varrho:\mathbb{C}\to \mathbb{C}$ that have the property that their associated networks are universal, i.e., are capable of approximating continuous functions to arbitrary accuracy on compact domains. Precisely, we show that deep narrow complex-valued networks are universal if and only if their activation function is neither holomorphic, nor antiholomorphic, nor $\mathbb{R}$-affine. This is a much larger class of functions than in the dual setting of arbitrary width and fixed depth. Unlike in the real case, the sufficient width differs significantly depending on the considered activation function. We show that a width of $2n+2m+5$ is always sufficient and that in general a width of $max\{2n,2m\}$ is necessary. We prove, however, that a width of $n+m+3$ suffices for a rich subclass of the admissible activation functions. Here, $n$ and $m$ denote the input and output dimensions of the considered networks. Moreover, for the case of smooth and non-polyharmonic activation functions, we provide a quantitative approximation bound in terms of the depth of the considered networks.