Restricting the Flow: Information Bottlenecks for Attribution
Restricting the Flow: Information Bottlenecks for Attribution
复制标题
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Karl Schulz;Leon Sixt;Federico Tombari;Tim Landgraf
中科院分区：
文献类型：
--
作者：
Karl Schulz;Leon Sixt;Federico Tombari;Tim Landgraf
归因方法为人工神经网络等机器学习模型的决策提供了见解。对于给定的输入样本，它们为每个输入变量（例如图像的像素）分配相关性得分。在这项工作中，我们适应信息瓶颈的概念归属。通过向中间特征图添加噪声，我们限制了信息流，并可以量化（以比特为单位）图像区域提供了多少信息。我们在VGG-16和ResNet-50上使用三种不同的指标将我们的方法与十个基线进行比较，发现我们的方法在六个设置中的五个设置中优于所有基线。该方法的信息理论基础为属性值（位）提供了一个绝对的参考框架，并保证网络决策不需要得分接近零的区域。
Attribution methods provide insights into the decision-making of machine learning models like artificial neural networks. For a given input sample, they assign a relevance score to each individual input variable, such as the pixels of an image. In this work we adapt the information bottleneck concept for attribution. By adding noise to intermediate feature maps we restrict the flow of information and can quantify (in bits) how much information image regions provide. We compare our method against ten baselines using three different metrics on VGG-16 and ResNet-50, and find that our methods outperform all baselines in five out of six settings. The method’s information-theoretic foundation provides an absolute frame of reference for attribution values (bits) and a guarantee that regions scored close to zero are not required for the network's decision.