{ "id": "2401.16594", "version": "v2", "published": "2024-01-29T21:51:27.000Z", "updated": "2024-06-20T22:47:03.000Z", "title": "Consistent algorithms for multi-label classification with macro-at-$k$ metrics", "authors": [ "Erik Schultheis", "Wojciech Kotłowski", "Marek Wydmuch", "Rohit Babbar", "Strom Borman", "Krzysztof Dembczyński" ], "comment": "This is the authors' version of the work accepted to ICLR 2024; the final version of the paper, errors and typos corrected, and minor modifications to improve clarity", "categories": [ "cs.LG" ], "abstract": "We consider the optimization of complex performance metrics in multi-label classification under the population utility framework. We mainly focus on metrics linearly decomposable into a sum of binary classification utilities applied separately to each label with an additional requirement of exactly $k$ labels predicted for each instance. These \"macro-at-$k$\" metrics possess desired properties for extreme classification problems with long tail labels. Unfortunately, the at-$k$ constraint couples the otherwise independent binary classification tasks, leading to a much more challenging optimization problem than standard macro-averages. We provide a statistical framework to study this problem, prove the existence and the form of the optimal classifier, and propose a statistically consistent and practical learning algorithm based on the Frank-Wolfe method. Interestingly, our main results concern even more general metrics being non-linear functions of label-wise confusion matrices. Empirical results provide evidence for the competitive performance of the proposed approach.", "revisions": [ { "version": "v2", "updated": "2024-06-20T22:47:03.000Z" } ], "analyses": { "keywords": [ "multi-label classification", "consistent algorithms", "independent binary classification tasks", "population utility framework", "long tail labels" ], "note": { "typesetting": "TeX", "pages": 0, "language": "en", "license": "arXiv", "status": "editable" } } }