… We propose sigmoidF1, a loss function that is an approximation of the macro … We introduce
sigmoidF1, an F1 score surrogate, with a sigmoid function acting as a surrogate thresholding …

… We experiment with different losses to optimize the multi-label classifier, including binary …
de Rijke, Sigmoidf1: A smooth f1 score surrogate loss for multilabel classification, arXiv preprint …

… We explore methods to solve the multi-label classification task posed by the GeoLifeCLEF
2024 … de Rijke, Sigmoidf1: A smooth f1 score surrogate loss for multilabel classification, arXiv …

… using a smooth F1score loss function for multilabel multiclass … a surrogate model on top
of numerical modeling results by … not mutually exclusive (multiclass multilabel classification). …

… Having as a starting point a simple multi-label classifier based … with BCE (binary cross entropy)
loss, the validation F1 score … -dimensional matrix, thus we compute the sigmoidF1 loss as …

… (negative results) for coresets for both MCC and F1 scores. … coreset construction guarantees
for F1 score and MCC. We … sigmoidF1: A Smooth F1 Score Surrogate Loss for Multilabel …

… estimated to maximize micro- or macro-averaged F1 scores. … For example, sigmoidF1 from
[1] creates smooth versions for … Multi-label text categorization with model combination based …

… loss function consists of the weighted focal loss and the loss of the F1 score; together, these
are used to correctly classify … L}_{F1}$| approximates the F1 score by using smooth surrogate …

… added the sigmoidF1 loss function [21], which together with the BCE … F1 score, it addresses
the non-differentiability issue of the F1 function, allowing its use in multi-label classification …

… In our case, since we have a multilabel classification problem, the precision for each label
… de Rijke, ‘‘SigmoidF1: A smooth F1 score surrogate loss for multilabel classification,’’ 2021, …