We thank all reviewers for their efforts in reviewing our paper, and for the helpful comments and suggestions. "Id like to see some statistical properties of these new loss functions, and how they are compared to the "The authors are encouraged to present more theoretical & empirical analysis on the robustness of these loss We provide strong empirical evidence of the efficacy of our method, i.e. for dropping the convexity in the We leave this up to future work. Reviewer 3: We will add the additional references you suggested. "There are many possibilities for combining heavy-tailed distributions and robust/consistent divergence. The computational cost is only negligibly larger than the cost of logistic regression.

We thank all the reviewers for their helpful comments and suggestions. Synthetic data is a powerful tool that is frequently employed in ML. First, thank you very much for the thorough and thoughtful review, especially about the related work. RE: related work We actually agree with your characterization of the literature and see it as complementary to ours. Perhaps our language was sometimes too strong in our attempt to highlight what we perceive as a problem.

Firstly, we thank all reviewers for the helpful comments and suggestions. We will add citations in Table 4. We haven't conducted experiments in language modeling and image density estimation Admittedly, modeling the intra-step correlation would require extra computation time. We will add this discussion in the revised version. We are not entirely sure about the motivation of the multi-frame setting.