Trustworthy machine learning in oncology | Florian Büttner

In this talk I will give an overview on some challenges and opportunities of developing machine learning approaches tailored for translational applications. I will present recent work on how to achieve (more) trustworthy predictions by ensuring calibration under domain shift. In addition, I will speak about how hierarchical autoencoders can facilitate unsupervised domain generalization. Finally, I will present recent work on multi-output Gaussian Process latent variable models as uncertainty-aware recommender systems.

https://www.dkfz.de/en/datascience/seminar/Buettner.html

Видео Trustworthy machine learning in oncology | Florian Büttner канала Intelligent Medical Systems