Machine Learning at Microsoft with ML .NET
Authors:
- Zeeshan AhmedZAZeeshan Ahmed
- Saeed AmizadehSASaeed Amizadeh
- Mikhail BilenkoMBMikhail Bilenko
- Rogan CarrRCRogan Carr
- Wei-Sheng ChinWCWei-Sheng Chin
- Yael DekelYDYael Dekel
- Xavier DupreXDXavier Dupre
- Vadim EksarevskiyVEVadim Eksarevskiy
- Eric ErhardtEEEric Erhardt
- Costin EseanuCECostin Eseanu
- Senja FilipiSFSenja Filipi
- Tom FinleyTFTom Finley
Preprint
Abstract
Machine Learning is transitioning from an art and science into a technology available to every developer. In the near future, every application on every platform will incorporate trained models to encode data-based decisions that would be impossible for developers to author. This presents a significant engineering challenge, since currently data science and modeling are largely decoupled from standard software development processes. This separation makes incorporating machine learning capabilities inside applications unnecessarily costly and difficult, and furthermore discourage developers from embracing ML in first place. In this paper we present ML .NET, a framework developed at Microsoft over the last decade in response to the challenge of making it easy to ship machine learning models in large software applications. We present its architecture, and illuminate the application demands that shaped it. Specifically, we introduce DataView, the core data abstraction of ML .NET which allows it to capture full predictive pipelines efficiently and consistently across training and inference lifecycles. We close the paper with a surprisingly favorable performance study of ML .NET compared to more recent entrants, and a discussion of some lessons learned.Keywords
All Related Versions
- Version 1, 2019-05-14, ArXiv
- Version 2, 2019-05-15, ArXiv
- Published version: , 2448.
This publication has 6 references indexed in Scilit:
- Scaling Up Stochastic Dual Coordinate Ascent2015
- The NumPy Array: A Structure for Efficient Numerical ComputationComputing in Science & Engineering, 2011
- Pegasos: primal estimated sub-gradient solver for SVMMathematical Programming, 2010
- Scalable training ofL1-regularized log-linear models2007
- Greedy function approximation: A gradient boosting machine.The Annals of Statistics, 2001
- Volcano-an extensible and parallel query evaluation systemIEEE Transactions on Knowledge and Data Engineering, 1994