Gaussian Compression Stream: Principle and Preliminary Results
This work addresses computational bottlenecks in big data processing for researchers and practitioners using NMF, though it appears incremental as it modifies existing compression methods.
The paper tackles the inefficiency of structured random projections in Nonnegative Matrix Factorization (NMF) by proposing Gaussian compression stream, an alternative based solely on Gaussian compressions that leverages fast techniques and is shown to be well-suited for NMF.
Random projections became popular tools to process big data. In particular, when applied to Nonnegative Matrix Factorization (NMF), it was shown that structured random projections were far more efficient than classical strategies based on Gaussian compression. However, they remain costly and might not fully benefit from recent fast random projection techniques. In this paper, we thus investigate an alternative to structured ran-om projections-named Gaussian compression stream-which (i) is based on Gaussian compressions only, (ii) can benefit from the above fast techniques, and (iii) is shown to be well-suited to NMF.