This article "Consistency for L2 Boosting and Matching Pursuit with Trees and Tree-type Basis Functions" from Buhlmann has some results on consistency for Boosting but they are for convergence in probability. I wonder if there are similar results for almost sure convergence.
The reason I am asking this is that I am doing some calculations that depend on these results but convergence in probability is not enough because at some point I need to evoke the dominated convergence theorem to exchange a limit with an integral and I can not do that unless I have almost sure results.
Thanks!