ABSTRACT
While Convolutional Neural Network (CNN) is known to be a prominent method, there is still a need for appropriate structures to achieve high accuracy. Structure optimization is usually handcrafted, and there is a raising need for an automated optimization method. In this research, we introduce an auto-encoder to create new structures and apply Differential Evolution with an Individual-Dependent Mechanism (IDE) to a simple CNN Structure optimization problem. Experiments using the proposed framework have been conducted on the Cifar10 dataset. Experimental results showed that the proposed method is fit as a structure optimizer.
- Diederik P Kingma and Max Welling. 2013. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114 (2013).Google Scholar
- L. Tang, Y. Dong, and J. Liu. 2015. Differential Evolution With an Individual-Dependent Mechanism. IEEE Transactions on Evolutionary Computation 19, 4 (2015), 560--574. Google ScholarDigital Library
- Steven R. Young, Derek C. Rose, Thomas P. Karnowski, Seung-Hwan Lim, and Robert M. Patton. 2015. Optimizing Deep Learning Hyper-Parameters through an Evolutionary Algorithm. Association for Computing Machinery. Google ScholarDigital Library
Index Terms
- CNN structure optimization using differential evolution with individual dependent mechanism
Recommendations
CNN architecture optimization using bio-inspired algorithms for breast cancer detection in infrared images
AbstractThe early detection of breast cancer is a vital factor when it comes to improving cure and recovery rates in patients. Among such early detection factors, one finds thermography, an imaging technique that demonstrates good potential as an early ...
Highlights- Use of GA for CNNs optimization hyperparameters and fully connected layers.
- Use of PSO for CNNs optimization.
- Optimized CNN architectures for breast cancer detection.
- Infrared imagery to support computer-aided diagnosis of ...
Adaptive Shrinkage Function Optimization by Differential Evolution
ADCOM '07: Proceedings of the 15th International Conference on Advanced Computing and CommunicationsIn this paper, a new wavelet shrinkage denoising algorithm is presented. The algorithm uses Wavelet Transform (WT) to extract information about sharp variation in multiresolution images and applies shrinkage function adapting the image features. The ...
Fast transformation of discriminators into encoders using pre-trained GANs
Highlights- A fast method for self-supervised discriminator-to-encoder transformation.
- The ...
AbstractA typical generative adversarial network (GAN) consists of a generator and a discriminator. Currently, finely tuned deep GANs can synthesize high-quality (HQ) images via their generators. However, the discriminator in typical GANs is ...
Comments