Learning Important Features Through Propagating Activation Differences
Avanti Shrikumar, Peyton Greenside, Anshul Bharat Kundaje · arXiv (Cornell University) · 2017
The purported "black box" nature of neural networks is a barrier to adoption in applications where interpretability is essential. Here we present DeepLIFT (Deep Learning Important FeaTures), a method for decomposing the output prediction of a neural network on a specific input by backpropagating the contributions of all neurons in the network to every feature of the input. DeepLIFT compares the activation of each neuron to its 'reference activation' and assigns contribution scores according to the difference. By optionally giving separate consideration to positive and negative contributions, DeepLIFT can also reveal dependencies which are missed by other approaches. Scores can be computed efficiently in a single backward pass. We apply DeepLIFT to models trained on MNIST and simulated genomic data, and show significant advantages over gradient-based methods. Video tutorial: http://goo.gl/qKb7pL, ICML slides: bit.ly/deeplifticmlslides, ICML talk: https://vimeo.com/238275076, code: http://goo.gl/RM8jvH.
Cited by 2,199 works. doi:10.48550/arxiv.1704.02685 Open-access PDF
Authors
References
- Visualizing and Understanding Convolutional Networks
- Striving for Simplicity: The All Convolutional Net
- On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation
- Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
- Predicting effects of noncoding variants with deep learning–based sequence model
- Not Just a Black Box: Learning Important Features Through Propagating Activation Differences
- Visualizing Deep Neural Network Decisions: Prediction Difference Analysis
- Systematic discovery and characterization of regulatory motifs in ENCODE TF binding experiments
- An unexpected unity among methods for interpreting model predictions
- Gradients of Counterfactuals