arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1312.6199cs.CVcs.LGcs.NE
收藏

神经网络的有趣特性

Intriguing properties of neural networks

Google Inc.(谷歌公司);New York University(纽约大学);University of Montreal(蒙特利尔大学);Facebook Inc.(脸书公司)

Christian Szegedy, Wojciech Zaremba, Ilya Sutskever, Joan Bruna, Dumitru Erhan, Ian Goodfellow, Rob Fergus

摘要:

深度神经网络是高度表达的模型,最近在语音和视觉识别任务上实现了最先进的性能。虽然他们的表现力是他们成功的原因,但这也导致他们学习到无法解释的解决方案,这些解决方案可能具有违反直觉的特性。在本文中,我们报告两个这样的属性。首先,我们发现,有没有个别的高层次单位和高层次单位的随机线性组合之间的区别,根据各种方法的单元分析。这表明,在神经网络的高层中,是空间而不是单个单元包含语义信息。其次,我们发现深度神经网络学习的输入-输出映射在很大程度上是不连续的。我们可以通过施加某种不可感知的扰动来使网络对图像进行错误分类,这种扰动是通过最大化网络的预测误差来发现的。此外,这些扰动的特定性质不是学习的随机伪影:相同的扰动可能导致在数据集的不同子集上训练的不同网络对相同的输入进行错误分类。

英文摘要:

Deep neural networks are highly expressive models that have recently achieved state of the art performance on speech and visual recognition tasks. While their expressiveness is the reason they succeed, it also causes them to learn uninterpretable solutions that could have counter-intuitive properties. In this paper we report two such properties. First, we find that there is no distinction between individual high level units and random linear combinations of high level units, according to various methods of unit analysis. It suggests that it is the space, rather than the individual units, that contains of the semantic information in the high layers of neural networks. Second, we find that deep neural networks learn input-output mappings that are fairly discontinuous to a significant extend. We can cause the network to misclassify an image by applying a certain imperceptible perturbation, which is found by maximizing the network's prediction error. In addition, the specific nature of these perturbations is not a random artifact of learning: the same perturbation can cause a different network, that was trained on a different subset of the dataset, to misclassify the same input.

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。

↑