Global Explanations of Neural Networks: Mapping the Landscape of Predictions
This addresses the problem of opaque neural network decisions for practitioners, offering incremental improvements in global explanation methods.
The paper tackles the lack of interpretability in neural networks by introducing GAM, a method for generating global attributions that explain prediction landscapes across subpopulations, demonstrating it matches known feature importances on simulated data, aligns with interpretable models on real data, and is intuitive in user studies.
A barrier to the wider adoption of neural networks is their lack of interpretability. While local explanation methods exist for one prediction, most global attributions still reduce neural network decisions to a single set of features. In response, we present an approach for generating global attributions called GAM, which explains the landscape of neural network predictions across subpopulations. GAM augments global explanations with the proportion of samples that each attribution best explains and specifies which samples are described by each attribution. Global explanations also have tunable granularity to detect more or fewer subpopulations. We demonstrate that GAM's global explanations 1) yield the known feature importances of simulated data, 2) match feature weights of interpretable statistical models on real data, and 3) are intuitive to practitioners through user studies. With more transparent predictions, GAM can help ensure neural network decisions are generated for the right reasons.