Congratulations to Dewan Fayzur on winning the Animal Behavior Challenge 2018!

Posted by

Our esteemed colleague Dewan Fayzur, data scientist, recently won the Animal Behavior Challenge (ABC) 2018, a CodaLab competition organized by the Symposium on Systems Science of Bio-Navigation, and sponsored by Technosmart.

The goal of the challenge was to predict the gender of streaked shearwaters from their GPS trajectory using Machine Learning. The bird colony was located in the small island of Awashima, Japan, and consisted of approximately 84,000 specimens. Thanks to data loggers attached to the legs of 631 birds, it was possible to determine that males and females use different cues and trajectories, meaning it’s possible to predict their gender from the way they navigate.

Using Machine Learning to predict bird gender
The blue trajectory is male, red is female.

You can read Fayzur’s winning submission at the Cornell University Library ArXiv to learn all the technicalities of the project. You can also find the project’s source code on Github.

As part of the award, Fayzur will be traveling to Kyoto and attending the Symposium on Systems Science of Bio-Navigation 2018, from 5-6 September. We wish him a well-deserved fantastic time!

PowerBI Tiles Pro Banner 99

Leave a Reply

Your email address will not be published.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