We thank all the reviewers for their helpful and positive feedback. We are happy that all the reviewers acknowledge the contributions of this work, in particular, a novel loss function that addresses a relevant practical problem, an efficient optimization scheme that scales to large datasets, and an experimental evaluation that shows encouraging results. Reviewer 2: - motivate the choice of the loss function further. We provide two possible directions to extend the motivation. First, the top-k error is used as the performance measure in certain benchmarks, e.g. in the ImageNet challenge in computer vision.

The top-$k$ error is often employed to evaluate performance for challenging classification tasks in computer vision as it is designed to compensate for ambiguity in ground truth labels. This practical success motivates our theoretical analysis of consistent top-$k$ classification. To this end, we define top-$k$ calibration as a necessary and sufficient condition for consistency, for bounded below loss functions. Unlike prior work, our analysis of top-$k$ calibration handles non-uniqueness of the predictor scores, and extends calibration to consistency -- providing a theoretically sound basis for analysis of this topic. Based on the top-$k$ calibration analysis, we propose a rich class of top-$k$ calibrated Bregman divergence surrogates. Our analysis continues by showing previously proposed hinge-like top-$k$ surrogate losses are not top-$k$ calibrated and thus inconsistent. On the other hand, we propose two new hinge-like losses, one which is similarly inconsistent, and one which is consistent. Our empirical results highlight theoretical claims, confirming our analysis of the consistency of these losses.

In order to push the performance on realistic computer vision tasks, the number of classes in modern benchmark datasets has significantly increased in recent years. This increase in the number of classes comes along with increased ambiguity between the class labels, raising the question if top-1 error is the right performance measure. In this paper, we provide an extensive comparison and evaluation of established multiclass methods comparing their top-k performance both from a practical as well as from a theoretical perspective. Moreover, we introduce novel top-k loss functions as modifications of the softmax and the multiclass SVM losses and provide efficient optimization schemes for them. In the experiments, we compare on various datasets all of the proposed and established methods for top-k error optimization. An interesting insight of this paper is that the softmax loss yields competitive top-k performance for all k simultaneously. For a specific top-k error, our new top-k losses lead typically to further improvements while being faster to train than the softmax.