Airbnb announces Aerosolve, an open-source machine learning software package

Home-rental company Airbnb today introduced a new open-source project called Aerosolve.

The new tool, announced at Airbnb’s 2015 OpenAir developer conference in San Francisco, powers new pricing tips for hosts, which was also announced today. Written mostly in the Java and Scala programming languages, Aerosolve can also more intelligently rank and order things like images.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1742726,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,dev,","session":"D"}']

“This library is meant to be used with sparse, interpretable features such as those that commonly occur in search (search keywords, filters) or pricing (number of rooms, location, price),” Airbnb wrote in a description of the project on GitHub. “It is not as interpretable with problems with very dense non-human interpretable features such as raw pixels or audio samples.”

Airbnb engineers contribute to several open-source projects, like Apache Mesos, and they’ve also released their own projects for the rest of the world to use. Earlier this year Airbnb open-sourced its Airpal web user interface built on top of the PrestoDB open-source database. Aerosolve is the next step in turning Airbnb into a major open-source player, on par with Facebook (which created PrestoDB), LinkedIn, Netflix, Twitter, and other companies.

There are other open-source machine-learning libraries, like PredictionIO, Apache Mahout, and H2O, but a tool working at scale at a company like Airbnb should gain considerable adoption.

Find out more about Aerosolve in the new blog post about it.

Today Airbnb also announced Airflow, a new open-source tool for creating and monitoring data pipelines. Learn more about Airflow in this blog post.

And there’s a new website where Airbnb is showing off all of its open-source projects: Airbnb.io.