[Adapting Deep RankNet for Personalized Search](https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/wsdm233-song.pdf) is a paper published in 2014 by Song, Wang and He from Microsoft Research. It is heavily beased on [Learning to rank using gradient descent](https://www.shortscience.org/paper?bibtexKey=conf/icml/BurgesSRLDHH05) (Burges et al from Microsoft, 2005).
They use a neural network with 5 hidden layers. They investigate regularization by trunkated gradient and limiting the depth of the back propagation.
## See also
* July 2015: [RankNet: A ranking retrospective](https://www.microsoft.com/en-us/research/blog/ranknet-a-ranking-retrospective/)