English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Data compression is a popular technique for improving the efficiency of dataprocessing workloads such as SQL queries and more recently, machine learning(ML) with classical batch gradient methods. But the efficacy of such ideas formini-batch stochastic gradient descent (MGD), arguably the workhorse algorithmof modern ML, is an open question. MGD's unique data access pattern rendersprior art, including those designed for batch gradient methods, less effective.We fill this crucial research gap by proposing a new lossless compressionscheme we call tuple-oriented compression (TOC) that is inspired by an unlikelysource, the string/text compression scheme Lempel-Ziv-Welch, but tailored toMGD in a way that preserves tuple boundaries within mini-batches. We thenpresent a suite of novel compressed matrix operation execution techniquestailored to the TOC compression scheme that operate directly over thecompressed data representation and avoid decompression overheads. An extensiveempirical evaluation with real-world datasets shows that TOC consistentlyachieves substantial compression ratios by up to 51x and reduces runtimes forMGD workloads by up to 10.2x in popular ML systems.

Tuple-oriented Compression for Large-scale Mini-batch Stochastic Gradient Descent