Lets keep it simple, Using simple architectures to outperform deeper and more complex architectures

By Seyyed Hossein HasanPour, Mohammad Rouhani, Mohammad Sabokrou. 2016

130 citations recorded by the source corpus.

Major winning Convolutional Neural Networks (CNNs), such as AlexNet, VGGNet, ResNet, GoogleNet, include tens to hundreds of millions of parameters, which impose considerable computation and memory overhead. This limits their practical use for training, optimization and memory efficiency. On the contrary, light-weight architectures, being proposed to address this issue, mainly suffer from low accuracy. These inefficiencies mostly stem from following an ad hoc procedure. We propose a simple architecture, called SimpleNet, based on a set of designing principles, with which we empirically show, a well-crafted yet simple and reasonably deep architecture can perform on par with deeper and more complex architectures. SimpleNet provides a good tradeoff between the computation/memory efficiency and the accuracy. Our simple 13-layer architecture outperforms most of the deeper and complex architectures to date such as VGGNet, ResNet, and GoogleNet on several well-known benchmarks while having 2 to 25 times fewer number of parameters and operations. This makes it very handy for embedded system or system with computational and memory limitations. We achieved state-of-the-art result on CIFAR10 outperforming several heavier architectures, near state of the art on MNIST and competitive results on CIFAR100 and SVHN. Models are made available at: this https URL

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Neurology, Model Architectures

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yahia Saeed Assiri
2. A. Helen Victoria
3. G. Maragatham

In Computer Vision and Pattern Recognition

1. Yahia Saeed Assiri
2. A. Helen Victoria
3. G. Maragatham

In Neurology

1. Yahia Saeed Assiri
2. A. Helen Victoria
3. G. Maragatham

In Model Architectures

1. Yahia Saeed Assiri
2. A. Helen Victoria
3. G. Maragatham

Citations in the corpus

• Stochastic Optimization of Plain Convolutional Neural Networks with Simple Methods (2020, methodology): “As a result of this concept, We built simple convolutional network models[9] (plain convolutional neural networks with no residual blocks ), four layers for MNIST dataset, and 11 layers for Cifar10, Cifar100, SVHN datasets and 13 layers for STL10, The networks employ a homogeneous design utilizing 3”
• Mish: A Self Regularized Non-Monotonic Activation Function (2020, unknown): “Architecture Mish Swish ReLU ResNet-20 [15] 92.02 % 91.61% 91.71% WRN-10-2 [50] 86.83 % 86.56% 84.56% SimpleNet [12] 91.70 % 91.44% 91.16% Xception Net [5] 88.73 % 88.56% 88.38% Capsule Net [40] 83.15 % 82.48% 82.19% Inception ResNet v2 [42] 85.21 % 84.96% 82.22% DenseNet-121 [19] 91.27 % 90.92%…”
• Automatic tuning of hyperparameters using Bayesian optimization (2020, methodology): “This shows that there exists a trade-off between the model training and hyperparameter tuning time, which is quite useful (Hasanpour et al. 2018) when the trained model is used for other domains as well.”