Deformable Butterfly: A Highly Structured and Sparse Linear Transform
This work addresses network compression for AI practitioners, offering an incremental improvement through a novel structured method.
The authors tackled the problem of neural network compression by introducing Deformable Butterfly (DeBut), a structured and sparse linear transform that replaces standard layers, achieving light weight and low inference complexity without accuracy loss.
We introduce a new kind of linear transform named Deformable Butterfly (DeBut) that generalizes the conventional butterfly matrices and can be adapted to various input-output dimensions. It inherits the fine-to-coarse-grained learnable hierarchy of traditional butterflies and when deployed to neural networks, the prominent structures and sparsity in a DeBut layer constitutes a new way for network compression. We apply DeBut as a drop-in replacement of standard fully connected and convolutional layers, and demonstrate its superiority in homogenizing a neural network and rendering it favorable properties such as light weight and low inference complexity, without compromising accuracy. The natural complexity-accuracy tradeoff arising from the myriad deformations of a DeBut layer also opens up new rooms for analytical and practical research. The codes and Appendix are publicly available at: https://github.com/ruilin0212/DeBut.