Practical Block-wise Neural Network Architecture Generation

TL;DR

This paper introduces BlockQNN, an automated, block-wise neural network architecture generation method using reinforcement learning, achieving competitive results efficiently and demonstrating strong generalizability across datasets.

Contribution

It presents a novel block-wise network generation pipeline using Q-Learning, significantly reducing search time and outperforming hand-crafted and existing auto-generated networks.

Findings

Achieves 3.54% top-1 error on CIFAR-10, surpassing existing auto-generated networks.

Reduces search time to 3 days with 32 GPUs.

Networks generalize well from CIFAR to ImageNet.

Abstract

Convolutional neural networks have gained a remarkable success in computer vision. However, most usable network architectures are hand-crafted and usually require expertise and elaborate design. In this paper, we provide a block-wise network generation pipeline called BlockQNN which automatically builds high-performance networks using the Q-Learning paradigm with epsilon-greedy exploration strategy. The optimal network block is constructed by the learning agent which is trained sequentially to choose component layers. We stack the block to construct the whole auto-generated network. To accelerate the generation process, we also propose a distributed asynchronous framework and an early stop strategy. The block-wise generation brings unique advantages: (1) it performs competitive results in comparison to the hand-crafted state-of-the-art networks on image classification, additionally, the best network generated by BlockQNN achieves 3.54% top-1 error rate on CIFAR-10 which beats all existing auto-generate networks. (2) in the meanwhile, it offers tremendous reduction of the search space in designing networks which only spends 3 days with 32 GPUs, and (3) moreover, it has strong generalizability that the network built on CIFAR also performs well on a larger-scale ImageNet dataset.