论文标题
颠倒进食神经网络的特征可视化过程
Inverting the Feature Visualization Process for Feedforward Neural Networks
论文作者
论文摘要
这项工作阐明了神经网络中特征可视化的可逆性。由于使用激活最大化通过特征可视化生成的输入通常不会产生其优化的特征目标，因此我们研究了对产生此输入的功能目标的优化。鉴于激活最大化中使用的目标函数可以测量给定输入类似于特征目标的程度，我们利用了此函数W.R.T.的梯度。输入是 - 最多可达到缩放因子 - - 目标中的线性。该观察结果用于通过计算最小化梯度的封闭形式解决方案来找到最佳特征目标。通过逆特征可视化，我们打算在网络敏感性上对某些认为具有目标而不是激活的输入的敏感性提供替代视图。
This work sheds light on the invertibility of feature visualization in neural networks. Since the input that is generated by feature visualization using activation maximization does, in general, not yield the feature objective it was optimized for, we investigate optimizing for the feature objective that yields this input. Given the objective function used in activation maximization that measures how closely a given input resembles the feature objective, we exploit that the gradient of this function w.r.t. inputs is---up to a scaling factor---linear in the objective. This observation is used to find the optimal feature objective via computing a closed form solution that minimizes the gradient. By means of Inverse Feature Visualization, we intend to provide an alternative view on a networks sensitivity to certain inputs that considers feature objectives rather than activations.
