Machine Learning in Today

Introbuction to Spark

Machine Learning on Spark

Large Scale Neural Network

Optimization

Machine Learning in Today

old Time : Before Mobile Internet Booming Era: Data Scale is small:

- Shallow analysis is enough, no need complex algorithm

- Limited data set could not get precise training model

- Computing Capability is limited by technology

Today: Big Data laid a foundation of Machine Learning:

the widespread of smart phones and the development of IoT provides comprehensive data sources for big data。

Data are the Greatest Strategic Resources for Internet Companies.

In the past couple of decades, Computing Power is growing exponentially by following the Moore‘s Law

With the rapid development of science and the technology, more Complex Models are extracted, built-up and deploy in industry

More and more Efficient Algorithms of Machined Learning are researched and developed by scientists and domain experts.

ML: Big data + Computing Power + Complex Model + Efficient Algorithm

Create user experiences

Create commercial values

Machine learning Phase -1

Machine learning Phase -1

Machine learning Phase -1

Machine learning Phase -1

本文出自 “運維自動化” 博客，請務必保留此出處http://shower.blog.51cto.com/4926872/1976234

Machine Learning on Spark