Name
Abstract | ||
|---|---|---|
While neural networks are powerful approximators used to classify or embed data into lower dimensional spaces, they are often regarded as black boxes with uninterpretable features. Here we propose Graph Spectral Regularization for making hidden layers more interpretable without significantly impacting performance on the primary task. Taking inspiration from spatial organization and localization of neuron activations in biological networks, we use a graph Laplacian penalty to structure the activations within a layer. This penalty encourages activations to be smooth either on a predetermined graph or on a feature-space graph learned from the data via co-activations of a hidden layer of the neural network. We show numerous uses for this additional structure including cluster indication and visualization in biological and image data sets. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1007/978-3-030-44584-3_40 | IDA |
Keywords | DocType | Volume |
Feature saliency,Graph learning,Neural Network Interpretability | Conference | 12080 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
9 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Alexander Tong | 1 | 0 | 1.69 |
| David van Dijk | 2 | 40 | 5.77 |
| Jay S. Stanley | 3 | 0 | 0.34 |
| Matthew Amodio | 4 | 5 | 2.44 |
| Kristina Yim | 5 | 0 | 0.34 |
| Rebecca Muhle | 6 | 0 | 0.34 |
| James Noonan | 7 | 0 | 0.34 |
| Guy Wolf | 8 | 4 | 11.55 |
| Smita Krishnaswamy | 9 | 1 | 3.06 |