Fetching the paper…

Matrix Compression via Randomized Low Rank and Low Precision Factorization · Around