2013arXiv (Cornell University)Open access

Consistent Bounded-Asynchronous Parameter Servers for Distributed ML

Jinliang Wei, Wei Dai, Abhimanu Kumar, Xun Zheng, Qirong Ho, Eric P. Xing

Open full text 6 citations

Abstract

In distributed ML applications, shared parameters are usually replicated among computing nodes to minimize network overhead. Therefore, proper consistency model must be carefully chosen to ensure algorithm's correctness and provide high throughput. Existing consistency models used in general-purpose databases and modern distributed ML systems are either too loose to guarantee correctness of the ML algorithms or too strict and thus fail to fully exploit the computing power of the underlying distributed system. Many ML algorithms fall into the category of \emph{iterative convergent algorithms} which start from a randomly chosen initial point and converge to optima by repeating iteratively a set of procedures. We've found that many such algorithms are to a bounded amount of inconsistency and still converge correctly. This property allows distributed ML to relax strict consistency models to improve system performance while theoretically guarantees algorithmic correctness. In this paper, we present several relaxed consistency models for asynchronous parallel computation and theoretically prove their algorithmic correctness. The proposed consistency models are implemented in a distributed parameter server and evaluated in the context of a popular ML application: topic modeling.

Open-access reader

About this research paper

What this paper is about

In distributed ML applications, shared parameters are usually replicated among computing nodes to minimize network overhead. Therefore, proper consistency model must be carefully chosen to ensure algorithm's correctness and provide high throughput. Existing consistency models used in general-purpose databases and modern distributed ML systems are either too loose to guarantee correctness of the ML algorithms or too strict and thus fail to fully exploit the computing power of the underlying distributed system. Many ML algorithms fall into the category of \emph{iterative convergent algorithms} which start from a randomly chosen initial point and converge to optima by repeating iteratively a set of procedures. We've found that many such algorithms are to a bounded amount of inconsistency and still converge correctly. This property allows distributed ML to relax strict consistency models to improve system performance while theoretically guarantees algorithmic correctness. In this paper, we present several relaxed consistency models for asynchronous parallel computation and theoretically prove their algorithmic correctness. The proposed consistency models are implemented in a distributed parameter server and evaluated in the context of a popular ML application: topic modeling.

Why it matters

OpenAlex reports 6 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

In distributed ML applications, shared parameters are usually replicated among computing nodes to minimize network overhead. Therefore, proper consistency model must be carefully chosen to ensure algorithm's correctness and provide high throughput. Existing consistency models used in general-purpose databases and modern distributed ML systems are either too loose to guarantee correctness of the ML algorithms or too strict and thus fail to fully exploit the computing power of the underlying distributed system. Many ML algorithms fall into the category of \emph{iterative convergent algorithms} which start from a randomly chosen initial point and converge to optima by repeating iteratively a set of procedures. We've found that many such algorithms are to a bounded amount of inconsistency and still converge correctly. This property allows distributed ML to relax strict consistency models to improve system performance while theoretically guarantees algorithmic correctness. In this paper, we present several relaxed consistency models for asynchronous parallel computation and theoretically prove their algorithmic correctness. The proposed consistency models are implemented in a distributed parameter server and evaluated in the context of a popular ML application: topic modeling.

Key concepts: Bounded function, Asynchronous communication, Server, Computer science, Distributed computing, Computer network, Mathematics, Mathematical analysis

Related papers

Back to paper searchBrowse research topicsOriginal source
Consistent Bounded-Asynchronous Parameter Servers for Distributed ML — Research Paper | ScholarLens