arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.02162cs.CV
收藏

UnrealNAS:我们能用虚幻的数据搜索神经架构吗?

UnrealNAS: Can We Search Neural Architectures with Unreal Data?

University of California, Berkeley(加州大学伯克利分校);King Abdullah University of Science and Technology(阿卜杜拉国王科技大学);Tsinghua University(清华大学);Peking University(北京大学)

Zhen Dong, Kaicheng Zhou, Guohao Li, Qiang Zhou, Mingfei Guo, Bernard Ghanem, Kurt Keutzer, Shanghang Zhang

摘要:

神经结构搜索(NAS)在深度神经网络(DNN)的自动设计中取得了巨大的成功。然而,使用数据搜索网络架构的最佳方式仍不清楚,正在探索中。以前的工作已经分析了在NAS中使用地面实况标签的必要性,并激发了广泛的兴趣。在这项工作中,我们进一步质疑NAS是否需要真实数据才能有效。这个问题的答案对于可访问数据量有限的应用程序非常重要,并且可以帮助人们通过利用数据生成的额外灵活性来改进NAS。为了探索NAS是否需要真实数据,我们构建了三种类型的非真实数据集:1)随机标记的真实图像; 2)生成的图像和标签;以及3)生成的带有随机标签的高斯噪声。这些数据集有助于分析所搜索的架构的泛化性和表达性。我们研究了这些构造的数据集上使用流行的可区分NAS方法搜索的架构的性能。在CIFAR、ImageNet和CheXpert上进行的大量实验表明,与传统的NAS流水线相比,所搜索的架构可以获得令人满意的结果,这表明了使用非真实数据执行NAS的可行性。

英文摘要:

Neural architecture search (NAS) has shown great success in the automatic design of deep neural networks (DNNs). However, the best way to use data to search network architectures is still unclear and under exploration. Previous work has analyzed the necessity of having ground-truth labels in NAS and inspired broad interest. In this work, we take a further step to question whether real data is necessary for NAS to be effective. The answer to this question is important for applications with limited amount of accessible data, and can help people improve NAS by leveraging the extra flexibility of data generation. To explore if NAS needs real data, we construct three types of unreal datasets using: 1) randomly labeled real images; 2) generated images and labels; and 3) generated Gaussian noise with random labels. These datasets facilitate to analyze the generalization and expressivity of the searched architectures. We study the performance of architectures searched on these constructed datasets using popular differentiable NAS methods. Extensive experiments on CIFAR, ImageNet and CheXpert show that the searched architectures can achieve promising results compared with those derived from the conventional NAS pipeline with real labeled data, suggesting the feasibility of performing NAS with unreal data.

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。

↑