Heatmaps generated on inputs of image classification networks via explainable AI methods like Grad-CAM and LRP have been observed to resemble segmentations of input images in many cases. Consequently, heatmaps have also been leveraged for achieving weakly supervised segmentation with image-level supervision. On the other hand, losses can be imposed on differentiable heatmaps, which has been shown to serve for (1)~improving heatmaps to be more human-interpretable, (2)~regularization of networks towards better generalization, (3)~training diverse ensembles of networks, and (4)~for explicitly ignoring confounding input features. Due to the latter use case, the paradigm of imposing losses on heatmaps is often referred to as "Right for the right reasons". We unify these two lines of research by investigating semi-supervised segmentation as a novel use case for the Right for the Right Reasons paradigm. First, we show formal parallels between differentiable heatmap architectures and standard encoder-decoder architectures for image segmentation. Second, we show that such differentiable heatmap architectures yield competitive results when trained with standard segmentation losses. Third, we show that such architectures allow for training with weak supervision in the form of image-level labels and small numbers of pixel-level labels, outperforming comparable encoder-decoder models. Code is available at \url{https://github.com/Kainmueller-Lab/TW-autoencoder}.

翻译：通过可解释AI方法（如Grad-CAM和LRP）在图像分类网络输入上生成的热力图，在许多情况下被观察到类似于输入图像的分割结果。因此，热力图也被用于实现基于图像级监督的弱监督分割。另一方面，可在可微分热力图上施加损失函数，这已被证明可用于：（1）改进热力图以使其更易于人类理解；（2）对网络进行正则化以提升泛化能力；（3）训练多样化的网络集成；（4）显式忽略输入中的混淆特征。基于最后一种用途，在热力图上施加损失函数的范式常被称为“为正确理由而正确”。本研究通过探索半监督分割作为“为正确理由而正确”范式的新应用场景，统一了这两个研究方向。首先，我们展示了可微分热力图架构与标准图像分割编码器-解码器架构之间的形式化对应关系。其次，我们证明此类可微分热力图架构在使用标准分割损失训练时能获得具有竞争力的结果。第三，我们证明此类架构能够以图像级标签和少量像素级标签形式的弱监督进行训练，其性能优于可比的编码器-解码器模型。代码发布于 \url{https://github.com/Kainmueller-Lab/TW-autoencoder}。