Machine Learning Research

Neural architecture search with reinforcement learning 주저자: Barret Zoph (Google Brain) the structure and connectivity of a  neural network can be typically specified by a variable-length string. It is therefore possible to use  a recurrent network  – the controller – to generate such string. a simple method of using a recurrent network to  generate convolutional architectures. the recurrent network can be trained with  a policy gradient method to maximize the expected accuracy of the sampled architectures GENERATE MODEL DESCRIPTIONS WITH A CONTROLLER RECURRENT NEURAL NETWORK we use a controller to generate architectural hyperparameters of neural networks. To be flexible, the controller is implemented as a recurrent neural network. Let’s suppose we would like to predict feedforward neural networks with only convolutional layers, we can use the controller to generate their hyperparameters as a sequence of tokens: In our experiments, the ...

논문제목: Learning Transferable Architectures for Scalable Image Recognition 주저자:  Barret Zoph ( Google Brain) Abstract a method to learn the model architectures  directly on the dataset of interest. we propose to search for  an architectural building block on a small dataset and then  transfer the block to a larger dataset. “NASNet search space” which enables transferability. In our experiments, we search for the best convolutional layer (or “cell”) on the CIFAR-10 dataset and then apply this cell to the ImageNet dataset by stacking together more copies of this cell, each with their own parameters to design a convolutional  architecture, which we name a “NASNet architecture”. We also introduce a new regularization technique called ScheduledDropPath that significantly improves generalization in the NASNet models. Although the cell is not searched for directly on ImageNet, a NASNet constructed from the best cell achieves, among the published works...