We present HUDD, a tool that supports safety analysis practices for systems enabled by Deep Neural Networks (DNNs) by automatically identifying the root causes for DNN errors and retraining the DNN. HUDD stands for Heatmap-based Unsupervised Debugging of DNNs, it automatically clusters error-inducing images whose results are due to common subsets of DNN neurons. The intent is for the generated clusters to group error-inducing images having common characteristics, that is, having a common root cause. HUDD identifies root causes by applying a clustering algorithm to matrices (i.e., heatmaps) capturing the relevance of every DNN neuron on the DNN outcome. Also, HUDD retrains DNNs with images that are automatically selected based on their relatedness to the identified image clusters. Our empirical evaluation with DNNs from the automotive domain have shown that HUDD automatically identifies all the distinct root causes of DNN errors, thus supporting safety analysis. Also, our retraining approach has shown to be more effective at improving DNN accuracy than existing approaches. A demo video of HUDD is available at https://youtu.be/drjVakP7jdU.
翻译:我们展示了HUDD, 这是一种支持深神经网络(DNN)驱动系统的安全分析做法的工具,它通过自动识别 DNN错误的根源和再培训 DNN。 HUDD 代表基于热马普的DNN的无监督调试,它自动组合出错诱导图像,其结果因DNN神经的普通子集而产生。生成的组群旨在将具有共同特性的、即有共同根源的错误诱导图像组合在一起。HUDDD通过对矩阵(即Heatmaps)应用组合算法,捕捉DNNN结果中每个DNNN神经的关联性,从而查明根源。此外,HUDD重新显示,根据与已确定的图像组群的关联性,自动选择了HUDD Retrains DNN。我们与汽车域的DNNN的经验评价显示,HUDD自动识别出DNN错误的所有明显的根源,从而支持安全分析。此外,我们的再培训方法显示,在提高DNNNN的准确性方面比现有方法更有效。HUDDDDDD的演示视频可在 https://vr/Ud.