ASYNC: a cloud engine with asynchrony and history for distributed machine learning

Saeed Soori; Bugra Can; Mert Gürbüzbalaban; Maryam Mehri Dehnavi

doi:10.1109/IPDPS47924.2020.00052

Conference paper, 2020

ASYNC: a cloud engine with asynchrony and history for distributed machine learning

Saeed Soori, Bugra Can, Mert Gürbüzbalaban, Maryam Mehri Dehnavi

IEEE International Parallel & Distributed Processing Symposium (IPDPS), pp. 429–439, 2020.

Publisher (DOI)

In brief

A cloud computing engine for distributed machine-learning algorithms that are asynchronous or that depend on the history of past updates — the two features that make such algorithms fast and that standard frameworks handle badly. ASYNC is built on Spark, exposes the bookkeeping the algorithms need, and is demonstrated on asynchronous gradient and quasi-Newton methods.

Topics

Distributed & decentralized learning

Cite

@inproceedings{soori2020async,
  title   = {ASYNC: a cloud engine with asynchrony and history for distributed machine learning},
  author  = {Saeed Soori and Bugra Can and Mert Gürbüzbalaban and Maryam Mehri Dehnavi},
  year    = {2020},
  booktitle = {IEEE International Parallel & Distributed Processing Symposium (IPDPS)},
  pages   = {429--439},
  doi     = {10.1109/IPDPS47924.2020.00052},
}