Deterministic binary filters for convolutional neural networks
We propose Deterministic Binary Filters, an approach to Convolutional Neural Networks that learns weighting coefficients of predefined orthogonal binary basis instead of the conventional approach of learning directly the convolutional filters. This approach results in model architectures with signif...
| Hlavní autoři: | , , , , , |
|---|---|
| Médium: | Conference item |
| Vydáno: |
International Joint Conferences on Artificial Intelligence Organization
2018
|