Though over 100 papers have been published in the learning to rank (LTR) field, most of the large-scale, real-world datasets are not publicly available. This makes drawing comparisons between algorithms difficult.
In the spirit of changing this, Yahoo! is hosting the Learning to Rank Challenge. We'll offer up two of our never before released actual datasets. These datasets—used for learning our search ranking function—can only be accessed through participation in the competition.
This exciting machine learning challenge will consist of two tracks: the first is a standard LTR track and the second is a transfer-learning track. Both tracks are open to all external research groups in academia and industry.
Here is the link