Fetching the paper…

Higher-order Count Sketch: Dimensionality Reduction That Retains Efficient Tensor Operations · Around