The idea of StackNet is inpsired by neural network, for learning to rank problems, we have RankNet and ListNet: https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/tr-2007-40.pdf
So is it possible to implement custom objective functions for StackNet so that we can implement a StackNetRanker? I also found that the support for ensembling for learning to rank problems is limited.