Si avvisa che
in data 21-09-2023, alle ore 14:30 precise
presso il Politecnico di Milano, in aula 16B.2.1,
nell’ambito delle attività del MOX, si svolgerà il seguente seminario:
Sesia Matteo, University of Southern California
Titolo: Adaptive conformal classification with noisy labels
Abstract: This work develops novel conformal prediction methods for classification tasks that can automatically adapt to random label contamination in the calibration sample, enabling more informative prediction sets with stronger coverage guarantees compared to state-of-the-art approaches. This is made possible by a precise theoretical characterization of the effective coverage inflation (or deflation) suffered by standard conformal inferences in the presence of label contamination, which is then made actionable through new calibration algorithms. Our solution is flexible and can leverage different modeling assumptions about the label contamination process, while requiring no knowledge about the data distribution or the inner workings of the machine-learning classifier. The advantages of the proposed methods are demonstrated through extensive simulations and an application to object classification with the CIFAR-10H image data set.
Reference: http://export.arxiv.org/abs/2309.05092
Il link per seguire il seminario online sarà reso disponibile pochi minuti prima dell’avvio del seminario al seguente Link: https://mox.polimi.it/mox-seminars/?id_evento=2315
Tutti gli interessati sono cordialmente invitati a partecipare, Laura Sangalli
—— Laura Maria Sangalli MOX - Dipartimento di Matematica Politecnico di Milano Piazza Leonardo da Vinci 32 20133 Milano - Italy (+39) 02 2399 4554 laura.sangalli@polimi.itmailto:laura.sangalli@polimi.it https://sangalli.faculty.polimi.it