Paul Larsen: Adversarial Regularization Regimes in Classification Tasks

V četrtek, 21. 3. 2024, ob 14:15 bo v predavalnici 3.06 v okviru seminarja VeSFiM potekalo predavanja Paula Larsena z naslovom: Adversarial regularization regimes in classification tasks.

Povzetek: We demonstrate the possibility of a trend reversal in binary classification tasks between the dataset and a classification score obtained from a trained model. This trend reversal occurs for certain choices of the regularization parameter for model training, namely, if the parameter is contained in what we call the adversarial regularization regime. For ridge regression, we give necessary and sufficient algebraic conditions on the dataset for the existence of an adversarial regularization regime. Moreover, our results provide a data science practitioner with a hands-on tool to avoid hyperparameter choices suffering from trend reversal. We furthermore present numerical results on adversarial regularization regimes for logistic regression. Finally, we draw connections to datasets exhibiting Simpson’s paradox, providing a natural source of adversarial datasets. The talk is based on a joint paper with Maximilian Wiesmann, Max Planck Institute for Mathematics in the Sciences, Leipzig, Germany.

Predavanje bo potekalo v živo, bo pa omogočen tudi prenos prek interneta. Povezava na videokonferenčni sistem ZOOM: https://uni-lj-si.zoom.us/j/95492285750 ID: 954 9228 5750

Vljudno vabljeni!