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 1905.12152cs.LGcs.AIstat.ML
收藏

一种通过健全性检查的简单显著性方法

A Simple Saliency Method That Passes the Sanity Checks

Arushi Gupta, Sanjeev Arora

摘要:

人们对“显着性方法”(也称为“归因方法”)非常感兴趣,它通过为输入中的每个特征/像素分配“分数”来为深度网络的决策提供“解释”。它们的设计通常涉及通过输出相对于输入的梯度进行信用分配。最近Adebayo等人[arXiv:1810.03292]质疑了许多这些方法的有效性,因为它们没有通过简单的"健全性检查",即测试当训练网络的层被随机化时,或者当网络使用随机标签重新训练时,分数是否会发生变化/消失。我们提出了一个简单的修复现有的显着性方法,帮助他们通过健全的检查,我们称之为“像素竞争”。这涉及到在分类任务中计算所有可能标签的显着性图,并使用它们之间的简单竞争来识别并从图中删除不太相关的像素。最简单的变体是"竞争梯度$\odot$输入(CGI)":它是高效的,不需要额外的训练,并且只使用输入和梯度。它提供了一些理论依据(特别是对于ReLU网络),并通过经验证明了它的性能。

英文摘要:

There is great interest in "saliency methods" (also called "attribution methods"), which give "explanations" for a deep net's decision, by assigning a "score" to each feature/pixel in the input. Their design usually involves credit-assignment via the gradient of the output with respect to input. Recently Adebayo et al. [arXiv:1810.03292] questioned the validity of many of these methods since they do not pass simple *sanity checks* which test whether the scores shift/vanish when layers of the trained net are randomized, or when the net is retrained using random labels for inputs. We propose a simple fix to existing saliency methods that helps them pass sanity checks, which we call "competition for pixels". This involves computing saliency maps for all possible labels in the classification task, and using a simple competition among them to identify and remove less relevant pixels from the map. The simplest variant of this is "Competitive Gradient $\odot$ Input (CGI)": it is efficient, requires no additional training, and uses only the input and gradient. Some theoretical justification is provided for it (especially for ReLU networks) and its performance is empirically demonstrated.

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。

↑