Yandex has announced that it is open sourcing its machine learning library. The library is called CatBoost, and is based on gradient boosting. The term 'gradient boosting' refers to a “form of machine learning that analyses a wide range of data inputs”, and training and developing models to maximise the accuracy of predictions. In a press release, Yandex said by using this system, CatBoost should be able to deliver “highly accurate results even in situations where there is relatively little data”. CatBoost will be implemented into multiple Yandex products over the next few months, as a replacement for MatrixNet. It is also set to be used at CERN, the home of the Large Hardon Collider, in order to provide greater accuracy of data.Â
