This repository contains code and data associated with the paper Superiority of quadratic over conventional neural networks for classification of Gaussian mixture data. For more detail and future work, please check Fenglei Fan's works and QuadraLib.
To enrich the diversity of artificial neurons, a type of quadratic neurons was proposed previously, where the inner product of inputs and weights is replaced by a quadratic operation. In this paper, we demonstrate the superiority of such quadratic neurons over conventional counterparts. For this purpose, we train such quadratic neural networks using an adapted backpropagation algorithm and perform a systematic comparison between quadratic and conventional neural networks for classificaiton of Gaussian mixture data, which is one of the most important machine learning tasks. Our results show that quadratic neural networks enjoy remarkably better efficacy and efficiency than conventional neural networks in this context, and potentially extendable to other relevant applications.
Artifcial neural networks, Quadratic neurons, Quadratic neural networks, Backpropagation, Classifcation, Gaussian mixture models
This research work was carried out between 2022 and 2023 in Professor Ge Wang's AI-based X-ray Imaging System (AXIS) Lab. I would like to express my sincere gratitude to Prof. Wang for his guidance and assistance. His collective expertise has greatly contributed to my research during this period.