Structuring Neural Networks for More Explainable Predictions
Structuring Neural Networks for More Explainable Predictions
复制标题
DOI：
10.1007/978-3-319-98131-4_5
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Laura Rieger;Pattarawat Chormai;G. Montavon;L. K. Hansen;K. Müller
中科院分区：
文献类型：
--
作者：
Laura Rieger;Pattarawat Chormai;G. Montavon;L. K. Hansen;K. Müller
机器学习算法，如神经网络，当它们的预测可以解释时，例如在输入变量方面，更有用。通常，较简单的模型比具有较高性能的较复杂模型更易于解释。在实践中，人们可以选择一个易于解释（可能预测性较低）的模型。另一种解决方案是直接解释原始的、高度预测的模型。在本章中，我们提出了一种中间立场的方法，在这种方法中，为了减少在解释中观察到的常见偏差，对原始神经网络架构进行了节俭的修改。我们的方法可以更好地在前馈网络中分离类，并且可以更好地识别递归神经网络中的相关时间步长。
Machine learning algorithms such as neural networks are more useful, when their predictions can be explained, e.g. in terms of input variables. Often simpler models are more interpretable than more complex models with higher performance. In practice, one can choose a readily interpretable (possibly less predictive) model. Another solution is to directly explain the original, highly predictive model. In this chapter, we present a middle-ground approach where the original neural network architecture is modified parsimoniously in order to reduce common biases observed in the explanations. Our approach leads to explanations that better separate classes in feed-forward networks, and that also better identify relevant time steps in recurrent neural networks.