Reinforcement learning for neural architecture search: A review

Deep neural networks are efficient and flexible models that perform well for a variety of tasks such as image, speech recognition and natural language understanding. In particular, convolutional neural networks (CNN) generate a keen interest among researchers in computer vision and more specifically in classification tasks. CNN architecture and related hyperparameters are generally correlated to the nature of the processed task as the network extracts complex and relevant characteristics allowing the optimal convergence. Designing such architectures requires significant human expertise, substantial computation time and does not always lead to the optimal network. Reinforcement learning (RL) has been extensively used in automating CNN models design generating notable advances and interesting results in the field. This work aims at reviewing and discussing the recent progress of RL methods in Neural Architecture Search task and the current challenges that still require further consideration.