Radon machines: effective parallelisation for machine learning

Kamp, Michael and Boley, Mario and Missura, Olana and Gärtner, Thomas (2017) Radon machines: effective parallelisation for machine learning. Advances in Neural Information Processing Systems, 30 . ISSN 1049-5258

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (350kB) | Preview

Abstract

We present a novel parallelisation scheme that simplifies the adaptation of learning algorithms to growing amounts of data as well as growing needs for accurate and confident predictions in critical applications. In contrast to other parallelisation techniques, it can be applied to a broad class of learning algorithms without further mathematical derivations and without writing dedicated code, while at the same time maintaining theoretical performance guarantees. Moreover, our parallelisation scheme is able to reduce the runtime of many learning algorithms to polylogarithmic time on quasi-polynomially many processing units. This is a significant step towards a general answer to an open question [21] on efficient parallelisation of machine learning algorithms in the sense of Nick’s Class (NC). The cost of this parallelisation is in the form of a larger sample complexity. Our empirical study confirms the potential of our parallelisation scheme with fixed numbers of processors and instances in realistic application scenarios.

Item Type: Article
Additional Information: Acknowlegement of acceptance for publication in journal. Radon Machines: Effective Parallelisation for Machine Learning. Advances in Neural Information Processing Systems 30 (NIPS 2017). 31st Annual Conference: Neural Information Processing Systems 2017 held 4-9 December 2017, Long Beach, California.
Schools/Departments: University of Nottingham, UK > Faculty of Science > School of Computer Science
Related URLs:
Depositing User: Eprints, Support
Date Deposited: 24 Nov 2017 11:12
Last Modified: 05 Dec 2017 07:49
URI: http://eprints.nottingham.ac.uk/id/eprint/48362

Actions (Archive Staff Only)

Edit View Edit View