no code implementations • 18 Apr 2024 • Bartlomiej Sobieski, Przemysław Biecek
Specifically, we discover that the latent space of Diffusion Autoencoders encodes the inference process of a given classifier in the form of global directions.
1 code implementation • 2 Apr 2024 • Krzysztof Jankowski, Bartlomiej Sobieski, Mateusz Kwiatkowski, Jakub Szulc, Michal Janik, Hubert Baniecki, Przemyslaw Biecek
Foundation models have emerged as pivotal tools, tackling many complex tasks through pre-training on vast datasets and subsequent fine-tuning for specific applications.
3 code implementations • 17 Mar 2023 • Hubert Baniecki, Bartlomiej Sobieski, Patryk Szatkowski, Przemyslaw Bombinski, Przemyslaw Biecek
Time-to-event prediction, e. g. cancer survival analysis or hospital length of stay, is a highly prominent machine learning task in medical and healthcare applications.