Gradient Smoothing is an efficient approach to reducing noise in gradient-based model explanation method. SmoothGrad adds Gaussian noise to mitigate much of these noise. However, the crucial hyper-parameter in this method, the variance $\sigma$ of Gaussian noise, is set manually or with heuristic approach. However, it results in the smoothed gradients still containing a certain amount of noise. In this paper, we aim to interpret SmoothGrad as a corollary of convolution, thereby re-understanding the gradient noise and the role of $\sigma$ from the perspective of confidence level. Furthermore, we propose an adaptive gradient smoothing method, AdaptGrad, based on these insights. Through comprehensive experiments, both qualitative and quantitative results demonstrate that AdaptGrad could effectively reduce almost all the noise in vanilla gradients compared with baselines methods. AdaptGrad is simple and universal, making it applicable for enhancing gradient-based interpretability methods for better visualization.

翻译：梯度平滑是一种有效降低基于梯度的模型解释方法中噪声的技术。SmoothGrad通过添加高斯噪声来显著抑制此类噪声，但其关键超参数——高斯噪声的方差$\sigma$——通常依赖手动设置或启发式方法确定，这导致平滑后的梯度仍残留一定噪声。本文旨在将SmoothGrad阐释为卷积运算的推论，进而从置信度视角重新理解梯度噪声及$\sigma$的作用机理。基于此，我们提出一种自适应梯度平滑方法AdaptGrad。综合定性与定量实验表明，与基线方法相比，AdaptGrad能有效消除原始梯度中的绝大部分噪声。该方法简洁通用，可广泛应用于增强基于梯度的可解释性方法，以获得更清晰的可视化效果。