Neural Network Attributions: A Causal Perspective
This addresses the need for interpretable AI by providing a causal attribution method, which is novel but incremental in the field of explainable AI.
The authors tackled the problem of attributing neural network outputs to input features by proposing a new method based on first principles of causality, viewing the network as a Structural Causal Model, and reported experimental results on simulated and real datasets showcasing its promise and usefulness.
We propose a new attribution method for neural networks developed using first principles of causality (to the best of our knowledge, the first such). The neural network architecture is viewed as a Structural Causal Model, and a methodology to compute the causal effect of each feature on the output is presented. With reasonable assumptions on the causal structure of the input data, we propose algorithms to efficiently compute the causal effects, as well as scale the approach to data with large dimensionality. We also show how this method can be used for recurrent neural networks. We report experimental results on both simulated and real datasets showcasing the promise and usefulness of the proposed algorithm.