Characterizing I/O in Machine Learning with MLPerf Storage
ACM SIGMOD Record · 21.11.2022 · 51 (3) · 47-48
DOI: 10.1145/3572751.3572765 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM SIGMOD Record, online erschienen am 21.11.2022. Verfasst von Oana Balmau. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 11 Literaturangaben.
Abstract (Originalsprache)
Data is the driving force behind machine learning (ML) algorithms. The way we ingest, store, and serve data can impact the performance of end-to-end training and inference significantly [11]. However, efficient storage and pre-processing of training data has received far less focus in ML compared to efforts in building specialized software frameworks and hardware accelerators. The amount of data that we produce is growing exponentially, making it expensive and difficult to keep entire training datasets in main memory. Increasingly, ML algorithms will need to access data from persistent storage in an efficient way.
Zitieren
APA
Balmau, O. (2022). Characterizing I/O in Machine Learning with MLPerf Storage. ACM SIGMOD Record, 51(3), 47-48. https://doi.org/10.1145/3572751.3572765
Vancouver
Balmau O. Characterizing I/O in Machine Learning with MLPerf Storage. ACM SIGMOD Record. 2022;51(3):47-48. doi:10.1145/3572751.3572765
BibTeX
@article{Balmau2022,
author = {Balmau, Oana},
title = {Characterizing I/O in Machine Learning with MLPerf Storage},
journal = {ACM SIGMOD Record},
year = {2022},
volume = {51},
number = {3},
pages = {47--48},
doi = {10.1145/3572751.3572765}
}
EUVIMED