Abstract: Convolutional neural networks (CNNs) are very resource intensive and consume a lot of computational power. The convolution operation itself is a very complex process. Hence this work deals ...

Abstract: Winograd’s algorithm has demonstrated its advantages in accelerating the inference of convolution neural networks. It reduces the number of multiplications in convolution and has achieved ...

ABSTRACT: The first error theory and bounds for Fast Matrix Multiplication based on the Strassen-Winograd algorithms (FastMMW) were formulated in the 70s. The theory ...

Initial prototype rounds were of O(n!) complexity, able to calculate a max of n=18 arrays before crashing. Final target complexity is O(~n^2.374), which is the current best known complexity in theory ...

I was recently mulling on FPGA implementations of short length Fast Fourier Transforms (roughly N=30 to 100) at work. My dilemma was I needed a potentially arbitrary number of these in a scalable ...