Orthogonal Convolutional Neural Networks
This work addresses performance limitations in deep learning models for computer vision, offering a parameter-free solution with broad applicability, though it is incremental as it builds on existing orthogonality concepts.
The paper tackles training instability and feature redundancy in convolutional neural networks by proposing an efficient orthogonal convolution method based on doubly block-Toeplitz matrices, which consistently outperforms kernel orthogonality alternatives across tasks like image classification and inpainting in supervised, semi-supervised, and unsupervised settings.
Deep convolutional neural networks are hindered by training instability and feature redundancy towards further performance improvement. A promising solution is to impose orthogonality on convolutional filters. We develop an efficient approach to impose filter orthogonality on a convolutional layer based on the doubly block-Toeplitz matrix representation of the convolutional kernel instead of using the common kernel orthogonality approach, which we show is only necessary but not sufficient for ensuring orthogonal convolutions. Our proposed orthogonal convolution requires no additional parameters and little computational overhead. This method consistently outperforms the kernel orthogonality alternative on a wide range of tasks such as image classification and inpainting under supervised, semi-supervised and unsupervised settings. Further, it learns more diverse and expressive features with better training stability, robustness, and generalization. Our code is publicly available at https://github.com/samaonline/Orthogonal-Convolutional-Neural-Networks.