Game On! Introducing Cortana Intelligence Competitions

204 views Leave a comment

Machine Learning algorithms powered by intelligent applications offer useful functions in a daily lives in ways we might not even be wakeful of. For instance, predictive analytics concede businesses to keep pivotal customers, assistance open lines and buildings to run some-more efficiently, and assistance us find cinema that we are expected to find intriguing. The ML margin has gained extensive traction and honour over a final decade, call Harvard Business Review to name a Data Scientist a sexiest pursuit of a 21st century.


To inspire new ML applications and encourage a colourful online community, Cortana team launched Cortana Intelligence Competitions, a gamification underline of Cortana Intelligence Suite, as good as their first foe Decoding Brain Signals (currently ended) and second competition Womens Health Risk Assessment (about to end on 10/1/2016). This height provides an discerning and fun sourroundings to file users’ information scholarship and analytics expertise, and a initial foe will concede we to have a possibility to minister to a critical margin of neuroscience to win prizes and recognition.

Competitions concede we to:

  • Explore singular information sets by participating in varying levels of competitions. These datasets are being expelled publically for a initial time, so don’t skip out.
  • Compete with information scholarship experts during a tellurian level. What improved approach to learn than some healthy competition? Can we build predictive models with a aloft correctness than a experts? We concede we to contention mixed solutions, so keep training and trying.
  • Advance a information scholarship field, and in many cases scholarship in general, by contributing your ideas and creativity toward suggestive hurdles and pity formula with other experts.
  • Win prizes value thousands of dollars or find yourself on a desired Top 10 open leaderboard. Here’s your event for celebrity and glory!


You do not need to be an consultant to compete. In fact, we assimilated Microsoft dual months ago with no credentials in information science, and I’ve already participated in several inner competitions. Our tutorials, videos, and information set descriptions make it easy for those with an seductiveness in analytics and information scholarship to compete.

You’ll be means to contention your initial foe entrance in 4 easy steps:

  1. Find a foe you’d like to attend in a Cortana Intelligence Gallery. Then click on the Enter Competition button to duplicate a Starter Experiment into your existent Azure ML workspace. You can emanate a giveaway workspace though a credit label by simply logging in with a current Microsoft comment or Office 365 account. Add your special salsa here regulating possibly built-in modules or by bringing your R/Python scripts directly, to urge on a aim opening metric.
  2. Create a Predictive Experiment with a lerned indication out of your Starter Experiment, afterwards adjust a contention and outlay schema of a web use to safeguard they heed to a selection from a Competition documentation.
  3. Deploy a web use out of your Predictive Experiment. Test your web use regulating a Test symbol or a Excel template automatically combined for we to safeguard it is operative properly.
  4. Submit your web use as a foe entry, and see your open measure in a Cortana Intelligence Gallery foe page. And applaud if we make into a leaderboard!

After we successfully contention an entry, we can go behind to a copied Starter Experiment, iterate, and refurbish your Predictive Experiment, refurbish a web service, and contention a new entry.

Azure ML Studio provides a abounding set of ML modules as good as information estimate modules in a accessible GUI for constructing experiments. It also allows gifted information scientists to move tradition R and/or Python scripts for local execution. The R and Python runtime in a Studio come with a abounding set of open source R/Python packages, and additional packages can be alien as book bundles and referenced in a scripts as well.

The Competition reserve a Starter Experiment that leverages many built-in modules in Azure ML, though we can select to reinstate them with R or Python scripts assembled outwardly and alien into a Azure ML sourroundings to emanate a current entrance for submission.

Azure ML Studio also has a built-in JuPyteR Notebook use for we to do freestyle information exploration. Of course, we can always download a datasets used in a Competition and try it locally in your favorite tool.

For a finish set of manners and features, greatfully check out competitions’ FAQ. Team behind a platform prepared an sparkling foe rising with a recover of this feature, and it is open for participation. Enter today.

Source: Cortana Intelligence and Machine Learning Blog