Analyze and Design Network Architectures by Recursion Formulas
This work addresses the challenge of systematic network architecture design for researchers and practitioners in deep learning, though it appears incremental as it builds upon existing ResNet frameworks.
The paper tackled the problem of designing neural network architectures by analyzing recursion formulas, proposing a methodology to generate new architectures and demonstrating it with an improved version of ResNet that showed significant performance improvements on CIFAR and ImageNet datasets.
The effectiveness of shortcut/skip-connection has been widely verified, which inspires massive explorations on neural architecture design. This work attempts to find an effective way to design new network architectures. It is discovered that the main difference between network architectures can be reflected in their recursion formulas. Based on this, a methodology is proposed to design novel network architectures from the perspective of mathematical formulas. Afterwards, a case study is provided to generate an improved architecture based on ResNet. Furthermore, the new architecture is compared with ResNet and then tested on ResNet-based networks. Massive experiments are conducted on CIFAR and ImageNet, which witnesses the significant performance improvements provided by the architecture.