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 1812.09744cs.CV
收藏

利用类相似性提高深度神经网络鲁棒性

Leveraging Class Similarity to Improve Deep Neural Network Robustness

University of Denver(丹佛大学);University of California, Berkeley(加州大学伯克利分校)

Pooran Singh Negi, David chan, Mohammad Mahoor

摘要:

传统上,人工神经网络(ANN)通过最小化所提供的地面实况增量分布(编码为独热向量)与ANN的预测softmax分布之间的交叉熵来训练。然而,似乎不能接受的惩罚网络之间的分类错误平等。将类“Automobile”与类“Truck”相混淆的惩罚应该小于将类“Automobile”与类“Donkey”相混淆的惩罚。为了避免这种表示问题,并学习网络中更清晰的分类边界,本文提出了一种交叉熵损失的变化,它不仅取决于样本类,而且还取决于以矩阵形式编码的类之间的数据驱动的先验“类相似性分布”。我们探索使用数据驱动方法学习类相似性分布,然后表明通过使用我们修改的相似性驱动损失进行训练,我们在多个架构和数据集上获得了略好的泛化性能,并在嘈杂的测试场景中获得了更好的性能。

英文摘要:

Traditionally artificial neural networks (ANNs) are trained by minimizing the cross-entropy between a provided groundtruth delta distribution (encoded as one-hot vector) and the ANN's predictive softmax distribution. It seems, however, unacceptable to penalize networks equally for missclassification between classes. Confusing the class "Automobile" with the class "Truck" should be penalized less than confusing the class "Automobile" with the class "Donkey". To avoid such representation issues and learn cleaner classification boundaries in the network, this paper presents a variation of cross-entropy loss which depends not only on the sample class but also on a data-driven prior "class-similarity distribution" across the classes encoded in a matrix form. We explore learning the class-similarity distribution using a datadriven method and then show that by training with our modified similarity-driven loss, we obtain slightly better generalization performance over multiple architectures and datasets as well as improved performance on noisy testing scenarios.

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。

↑