New position available: Doctoral student in deep generative models for cancer research

Project description

Third-cycle subject: Computer Science

Variational autoencoders (VAEs), a part of the Variational inference (VI) methodology, has recently become important in analysis of large single-cell and spatial bio data. Yet, employing them for posterior approximations in Bayesian phylogenetics is a difficult problem. Clear examples of successful applications are still outstanding. Following recent trends in VAE research, one may hypothesize that it may be addressed by incorporating inductive biases, modeling more advanced prior distributions, and/or via architecture design.

This project aims to explore the possibility of employing VAEs or other deep generative models in the setting of somatic cancer evolution and Bayesian phylogenetics. By doing so, the Ph.D. will simultaneously contribute to two highly impactful scientific fields: computational bio/cancer and Bayesian inference in deep learning.

Our group teaches and produces cutting-edge research in VAEs, VI and approximate Bayesian inference in general, especially for methods in phylogenetics. The group provides a fertile environment for the Ph.D.’s development. The group also has a national and international network including both methodological and cancer-research groups.

Jens Lagergren is proposed to supervise the doctoral student. Decisions are made on admission

Last application date: 14.Oct.2022 11:59 PM CEST

Read more: