Metronome is a suite of parallel iterative algorithms that run natively on Hadoop's Next Generation YARN platform.

Algorithms
- Parallel Neural Networks
- Parallel linear regression
- Parallel logistic regression
- Parallel K-means
- Restricted Boltzmann Machines
- Parallel Deep Belief Networks
  - Developed directly in concert with [Adam Gibson and the DL4J project] (http://deeplearning4j.org)
  - [Running Deep Learning with Metronome on Hadoop] (https://github.com/jpatanooga/Metronome/wiki/Running-Deep-Learning-on-Metronome)
  - [Fundamental Architecture of our Deep Learning Library over at DL4J] (http://deeplearning4j.org/overview.html)
Scales linearly with input size
Built on top of BSP-style computation framework "Iterative Reduce" (Hadoop / YARN)
Packaged in a new suite of parallel iterative algorithms called Metronome on [IterativeReduce] (https://github.com/emsixteeen/IterativeReduce)
- 100% Java, ASF 2.0 Licensed, on github

Project Next Steps

6/5/2014 - we're working on cutting a 0.1 tag and download for the current project snapshot for both
- IterativeReduce
- Metronome

Presentations

Hadoop Summit EU 2013

[Hadoop Summit EU 2013: Parallel Linear Regression, IterativeReduce, and YARN] (http://www.slideshare.net/jpatanooga/hadoop-summit-eu-2013-parallel-linear-regression-iterativereduce-and-yarn)
[Video of Presentation from Hadoop Summit EU 2013] (http://www.youtube.com/watch?feature=player_embedded&v=rwiy_YhbxiI)

BigDataTechCon 2013

MLConf 2013

[Parallel Neural Networks on Hadoop Slides] (http://www.slideshare.net/jpatanooga/metronome-ml-confnov2013v20131113)

Hadoop Summit 2014 (San Jose)

[Slides for Intro To Deep Learning on Hadoop] (http://www.slideshare.net/jpatanooga/hadoop-summit-2014-san-jose-introduction-to-deep-learning-on-hadoop)
[Video of Hadoop Summit Talk] (https://www.youtube.com/watch?v=h2Ixpfn-DTg&list=PLSAiKuajRe2k0Aw_njj74rIlbdk3JTQ8Y)

Deep Learning Coverage in Wired Magazine

[Wired - Google Data Center AI] (http://www.wired.com/2014/05/google-data-center-ai/)
[Wired - Skymind Deep Learning] (http://www.wired.com/2014/06/skymind-deep-learning/)

[General guide on running yarn jobs] (https://github.com/jpatanooga/Metronome/wiki/Running-Jobs-on-YARN-Clusters)
[Running Deep Learning with Metronome on Hadoop] (https://github.com/jpatanooga/Metronome/wiki/Running-Deep-Learning-on-Metronome)
[IterativeReduce Programming Model] (https://github.com/emsixteeen/IterativeReduce/wiki/Iterative-Reduce-Programming-Guide)
Using IRUnit - the IterativeReduce Unit Testing Framework
[Running parallel linear regression with IRUnit on synthetic test data] (https://github.com/jpatanooga/Metronome/wiki/Running-Parallel-Linear-Regression)