All posts by
Aleksandra Osowska-Kurczab

1 year ago

Trust no one, not even your training data! Machine learning from noisy data

Label noise is ever-present in machine learning practice. Allegro datasets are no exception. We compared 7 methods for training classifiers robust to label noise. All of them improved the model’s performance on noisy datasets. Some of the methods decreased the model’s performance in the absence of label noise.

Aleksandra Osowska-Kurczab

Senior Research Engineer in the Machine Learning Research team, working on large-scale recommendation systems. She’s pursuing her PhD in Computer Science at Warsaw University of Technology on deep learning in medical image analysis. AI enthusiast and geek.