[Learning to rank using gradient descent](https://icml.cc/2015/wp-content/uploads/2015/06/icml_ranking.pdf) is a paper published in 2005 by Burges et all from Microsoft. The paper introduced RankNet.
RankNet is a neural network for recommendations.
The main use-case of the paper is ranking search results.
## Key Ideas
* Preprocessing: Filter results which are relevant
* Ranking: Rank results which are relevant by RankNet
## See also
* [Adapting deep RankNet for personalized search](https://www.shortscience.org/paper?bibtexKey=conf/wsdm/SongWH14)