[Question] Parallel implements for LASSO

Hi guys,

I wonder if there exists any MapReduce-based implement for LASSO, instead of MPI-based solutions? Or if I use bagging to combine several LARS results, will it be a good idea for large-scale data?

Thanks.