8767bccb1ff4231a9962e3914f4f1f8f-AuthorFeedback.pdf

A1: Thanks for your valued and constructive comments. The main1 motivation of our manuscript is to propose robust additive models (with theoretical analysis and applications) for2 realizing nonlinear estimation and structure discovery simultaneously, even data contaminated with complex noise3 and without priori knowledge of variable structure. The flexible selection ofν in our Outer Problem can unveil main effect variables across all8 tasks, which is useful to remove ambiguity for model identifiability. A3: Thanks for your valued suggestions. A6: [-When...]: mGAM with an oracle variable structure is the baseline of MAM.33