1 of 5

Machine learning at scale

​

Starts in July 14th, 2014

2 of 5

Need

  • fast algorithms for large data sets
    • recommendations
    • data exploration

​

3 of 5

Approach

  • faster k-NN
    • use clustering to speed up k-NN
    • using distributed RAM

​

4 of 5

Benefits

  • scalable algorithms
  • distributed
  • easy to deploy
  • easy to develop
  • applications
    • movie recommendation

5 of 5

Topics

  1. k-NN for HUGE datasets
  2. client-server clustering