Learning Fast Algorithms for Linear Transforms Using Butterfly Factorizations
This provides a method to automatically generate efficient and compressible transformations for machine learning pipelines, reducing the need for hand-crafted algorithms.
The paper tackles the problem of automatically learning fast algorithms for structured linear transforms, such as the FFT, by introducing a parameterization of divide-and-conquer methods; it recovers the O(N log N) Cooley-Tukey FFT to machine precision for N up to 1024 and, when used to compress a neural network, achieves 3.9 points higher accuracy on CIFAR-10 with 4x faster inference and 40x fewer parameters.
Fast linear transforms are ubiquitous in machine learning, including the discrete Fourier transform, discrete cosine transform, and other structured transformations such as convolutions. All of these transforms can be represented by dense matrix-vector multiplication, yet each has a specialized and highly efficient (subquadratic) algorithm. We ask to what extent hand-crafting these algorithms and implementations is necessary, what structural priors they encode, and how much knowledge is required to automatically learn a fast algorithm for a provided structured transform. Motivated by a characterization of fast matrix-vector multiplication as products of sparse matrices, we introduce a parameterization of divide-and-conquer methods that is capable of representing a large class of transforms. This generic formulation can automatically learn an efficient algorithm for many important transforms; for example, it recovers the $O(N \log N)$ Cooley-Tukey FFT algorithm to machine precision, for dimensions $N$ up to $1024$. Furthermore, our method can be incorporated as a lightweight replacement of generic matrices in machine learning pipelines to learn efficient and compressible transformations. On a standard task of compressing a single hidden-layer network, our method exceeds the classification accuracy of unconstrained matrices on CIFAR-10 by 3.9 points -- the first time a structured approach has done so -- with 4X faster inference speed and 40X fewer parameters.