Abstract:
The majority of existing neural networks operate with real-valued representation of data. However, there are multiple tasks in which the input is complex-valued. The comp...Show MoreMetadata
Abstract:
The majority of existing neural networks operate with real-valued representation of data. However, there are multiple tasks in which the input is complex-valued. The complex-valued data is considered to be more informative in terms of larger representational capacity. These reasons motivate researchers to develop neural networks using complex numbers instead of real-valued ones. In this paper, we take a step forward in the generalization of neural networks. We develop the basic building blocks for dual-valued neural networks based on dual numbers. We adjust basic layers such as Linear, Convolution, Average Pooling, ReLU to the dual domain and present an algorithm for Dual Batch Normalization. We construct several dual-valued neural networks for classification tasks basing on classical CV problems and the MusicNet and G2Net datasets. We show that dual-valued models outperform analogous complex-valued neural networks in execution time and have higher or at least the same accuracy.
Published in: 2022 18th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)
Date of Conference: 29 November 2022 - 02 December 2022
Date Added to IEEE Xplore: 24 November 2022
ISBN Information: