arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2103.02126cs.LG
收藏

基于可微神经结构学习的高效神经网络设计

Differentiable Neural Architecture Learning for Efficient Neural Network Design

Jiangnan University(江南大学);University of Jinan(济南大学);University of Surrey(萨里大学)

Qingbei Guo, Xiao-Jun Wu, Josef Kittler, Zhiquan Feng

摘要:

随着深度卷积神经网络(CNN)的发展,自动化神经网络设计受到了越来越多的关注,尤其是涉及它们在嵌入式和移动平台上的部署。神经架构搜索(NAS)面临的最大问题之一是需要大量候选神经架构来进行训练,例如使用强化学习和进化优化算法,这需要巨大的计算成本。即使是最近的可微分神经架构搜索(DNAS)也基于学习的架构参数的概率分布对少量候选神经架构进行采样,以选择最终的神经架构。为了解决这个计算复杂性的问题,我们引入了一种新的\emph{architecture parameterisation}基于缩放sigmoid函数,并提出了一个通用的\emph{Differentiable Neural Architecture Learning}(DNAL)方法来优化神经网络结构,而不需要评估候选神经网络。具体来说,对于随机超网以及传统的CNN,我们构建了一个新的通道模块层,其架构组件由缩放的sigmoid函数控制。我们从头开始训练这些神经网络模型。将网络优化解耦为权值优化和结构优化。我们解决了非凸优化问题的神经结构的连续缩放sigmoid方法的收敛保证。大量的实验表明,我们的DNAL方法在神经结构搜索成本方面提供了卓越的性能。通过DNAL学习的最佳网络在准确性、模型大小和计算复杂性方面超过了在基准CIFAR-10和ImageNet-1 K数据集上使用最先进方法生成的网络。

英文摘要:

Automated neural network design has received ever-increasing attention with the evolution of deep convolutional neural networks (CNNs), especially involving their deployment on embedded and mobile platforms. One of the biggest problems that neural architecture search (NAS) confronts is that a large number of candidate neural architectures are required to train, using, for instance, reinforcement learning and evolutionary optimisation algorithms, at a vast computation cost. Even recent differentiable neural architecture search (DNAS) samples a small number of candidate neural architectures based on the probability distribution of learned architecture parameters to select the final neural architecture. To address this computational complexity issue, we introduce a novel \emph{architecture parameterisation} based on scaled sigmoid function, and propose a general \emph{Differentiable Neural Architecture Learning} (DNAL) method to optimize the neural architecture without the need to evaluate candidate neural networks. Specifically, for stochastic supernets as well as conventional CNNs, we build a new channel-wise module layer with the architecture components controlled by a scaled sigmoid function. We train these neural network models from scratch. The network optimization is decoupled into the weight optimization and the architecture optimization. We address the non-convex optimization problem of neural architecture by the continuous scaled sigmoid method with convergence guarantees. Extensive experiments demonstrate our DNAL method delivers superior performance in terms of neural architecture search cost. The optimal networks learned by DNAL surpass those produced by the state-of-the-art methods on the benchmark CIFAR-10 and ImageNet-1K dataset in accuracy, model size and computational complexity.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑