Online Marketing Tools

Inside Google New algorithm RankBrain Data Center

Posted

Google_BrainRankBy Mary

 

BrainRank Machine learning. Google explains the machine like speech recognition (MI): a computer that analyzes large amounts of data to recognize patterns and “learn” from them. In a dynamic environment like Google’s  data center, it can be difficult for humans to see how all of the variables—IT load, outside air temperature, and the link.—interact with each other.

One thing computers are good at is seeing the underlying story in the data, so the information we gather in Google  daily operations and ran it through a model to help make sense of complex interactions that the team—being mere mortals—may not otherwise have noticed.

To keep the BrainRank machine, who wants to replace humans, running at its best. Google explains this simplified version of what the models do: take a couple of data, find the hidden interactions, then provide recommendations that optimize for  efficiency.

Google PUE Optimization Application

Google Use case: Predicting Power Usage Effectiveness (PUE)

-Basically: They developed a neural network framework that learns from operational data and models plant performance

–The model is able to predict PUE2 within a range of 0.004 + 0.005 , or 0.4% error for a PUE of 1.1.

google_AI

  • “A simplified version of what the models do: take a bunch of data, find the hidden interactions, then provide recommendations that optimize for energy efficiency.”

http://googleblog.blogspot.com/2014/05/better-data-centers-through-machine.html

Google Use Case: Features

  • Number of features relatively small (n = 19)Google_AI_RankBrain

Google Use Case: Algorithm

1.Randomly initialize the model parameters θ

2.Implement forward propagation

3.Compute the cost function J(θ)

4.Implement the back propagation algorithm

5.Repeat steps 2-4 until convergence

–or for the desired number of iterations

  • Very standard…

Google Use Case: Details

  • Neural Network

–5 hidden layers

–50 nodes per hidden layer

– 0.001 as the regularization parameter (λ)

  • Training Dataset

–19 normalized input parameters (features) per normalized output variable (the DC PUE)

  • Data normalized into the range [-1,-1]

Google_RankBrain_AI

– 184,435 time samples at 5 minute resolution

  • O(2) years of data

–70% for training, 30% for cross validation

Google_Brainrank

Google_IM

 

 

If you want to learn more about what is the new classical algorithm google search launched recently, please check this article: What is Google BrainRank?
If you want to take a 360 tour of Google’s Data Center make sure to use your Google Cardboard for better viewing expeirence.

Total Page Visits: 5863 - Today Page Visits: 1

Leave a Reply

Your email address will not be published. Required fields are marked *