Deep Learning with Noisy Labels

Explore the challenges and solutions for training deep learning models with noisy labels in this seminar by Professor Gustavo Carneiro from the University of Adelaide. Delve into the impact of mislabeled data on model performance and generalization, particularly in datasets collected from the internet or annotated by non-specialists. Discover various approaches proposed in literature to improve deep learning training in the presence of noisy labels, including new ideas developed by Professor Carneiro's research group. Gain insights into this crucial aspect of machine learning, especially relevant for fields like medical image analysis where label correctness can be challenging to guarantee.

[Seminar Series] Deep Learning with Noisy Labels