Quaternion-Valued Recurrent Projection Neural Networks on Unit Quaternions
This work addresses a specific issue in hypercomplex-valued neural networks for associative memory applications, representing an incremental improvement.
The authors tackled the cross-talk problem in quaternion-valued recurrent correlation neural networks (QRCNNs) by introducing quaternion-valued recurrent projection neural networks (QRPNNs), which combine non-local projection learning with QRCNNs, resulting in greater storage capacity and noise tolerance.
Hypercomplex-valued neural networks, including quaternion-valued neural networks, can treat multi-dimensional data as a single entity. In this paper, we present the quaternion-valued recurrent projection neural networks (QRPNNs). Briefly, QRPNNs are obtained by combining the non-local projection learning with the quaternion-valued recurrent correlation neural network (QRCNNs). We show that QRPNNs overcome the cross-talk problem of QRCNNs. Thus, they are appropriate to implement associative memories. Furthermore, computational experiments reveal that QRPNNs exhibit greater storage capacity and noise tolerance than their corresponding QRCNNs.