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 1804.04438cs.CVcs.LGstat.ML
收藏

池化对于CNN中适当的变形稳定性既不是必要的,也不是充分的

Pooling is neither necessary nor sufficient for appropriate deformation stability in CNNs

DeepMind(深度思维)

Avraham Ruderman, Neil C. Rabinowitz, Ari S. Morcos, Daniel Zoran

摘要:

我们关于神经网络如何运行的许多核心假设仍然没有经过经验检验。一个常见的假设是,卷积神经网络需要对小的平移和变形保持稳定,以解决图像识别任务。多年来,这种稳定性通过合并交错池化层而融入CNN架构中。然而,最近,交错池在很大程度上被放弃了。这就提出了一些问题:我们关于变形稳定性的直觉是否正确?很重要吗?变形不变性需要池化吗?如果不是,那么在没有变形不变性的情况下如何实现变形不变性?在这项工作中,我们严格测试了这些问题,并发现卷积网络中的变形稳定性比它最初出现的更微妙:(1)变形不变性不是一个二元属性,而是不同的任务需要不同程度的变形稳定性。(2)变形稳定性不是网络的固定属性,并且在训练过程中进行了大量调整,主要是通过卷积滤波器的平滑度。(3)交错池化层对于实现自然图像分类的变形稳定性的最佳形式既不是必要的,也不是充分的。(4)池化在初始化时为图像分类赋予了太多的变形稳定性,并且在训练期间,网络必须学习抵消这种归纳偏差。总之,这些发现为CNN中交织池化和变形不变性的作用提供了新的见解,并证明了对我们关于神经网络工作的最基本假设进行严格经验测试的重要性。

英文摘要:

Many of our core assumptions about how neural networks operate remain empirically untested. One common assumption is that convolutional neural networks need to be stable to small translations and deformations to solve image recognition tasks. For many years, this stability was baked into CNN architectures by incorporating interleaved pooling layers. Recently, however, interleaved pooling has largely been abandoned. This raises a number of questions: Are our intuitions about deformation stability right at all? Is it important? Is pooling necessary for deformation invariance? If not, how is deformation invariance achieved in its absence? In this work, we rigorously test these questions, and find that deformation stability in convolutional networks is more nuanced than it first appears: (1) Deformation invariance is not a binary property, but rather that different tasks require different degrees of deformation stability at different layers. (2) Deformation stability is not a fixed property of a network and is heavily adjusted over the course of training, largely through the smoothness of the convolutional filters. (3) Interleaved pooling layers are neither necessary nor sufficient for achieving the optimal form of deformation stability for natural image classification. (4) Pooling confers too much deformation stability for image classification at initialization, and during training, networks have to learn to counteract this inductive bias. Together, these findings provide new insights into the role of interleaved pooling and deformation invariance in CNNs, and demonstrate the importance of rigorous empirical testing of even our most basic assumptions about the working of neural networks.

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。

↑