Bootstrapping Machine Learning: overview and case study

Mar 5

Last Thursday I gave a Bootstrapping Machine Learning talk in which I explained how I reimplemented Gmail’s Priority Inbox. I have embedded the slides below, in which you'll find:

an explanation of the two phases of all Machine Learning systems;
snippets of Python and JS code employing Prediction APIs;
the format of the data I used to learn a model of email importance;
some methods of Google Apps Script you can use yourself to collect your own data on Google services;
a link with some more detailed information and example code;
links to my offline email analysis with BigML (dataset, model, and evaluation);
a link to a Google Spreadsheet that uses Google Apps Script and Google Prediction API to allow you to make predictions and fill in missing values in your spreadsheet.

All it takes to reimplement #Gmail’s #PriorityInbox is past #email #data and a #Prediction #API: http://t.co/NcghfxOsb5 #MachineLearning
— Louis Dorard (@louisdorard) March 2, 2014

This was my longest talk on that topic, and it’s also the one I’m most proud of! The format at BordeauxJS is very informal. I was half presenting and chatting with the audience, which is great because they had many pertinent questions to ask! I also really enjoyed seeing them realize the possibilities of ML as we were chatting. One of them actually told me that he finally had a solution for adding a long-wanted feature to his app, thanks to Prediction APIs!

If you have any questions regarding the slides above, let me know in the comments below!

Enjoyed this article? Follow me on Twitter!