Probabilistic models and deep learning - bridging the gap
Research Project, 2019 – 2024

In this project, we will develop theory and methods related to the interplay between probabilistic models and deep learning. More specifically, we intend to develop both new models and new inference and learning algorithms for applications where latent variables are naturally characterized using, for instance, probabilistic graphical models or stochastic processes, whereas data is from some domain where deep learning has been successful (e.g., images).

The family of problems that involve such interplay between deep learning and probabilistic models is general, and we expect to derive tools that are widely applicable. However, to make the research more concrete and to showcase the merits of the new methodology we will study three specific applications in more depth, each one of significant importance on its own: learning from weak annotations, dynamical systems with deep-learning-based likelihoods, and automated mitosis detection and counting in histopathology.

Participants

Lennart Svensson (contact)

Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering

Jakob Lindqvist

Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering

Collaborations

Linköping University

Linköping, Sweden

Funding

Wallenberg AI, Autonomous Systems and Software Program

(Funding period missing)

Related Areas of Advance and Infrastructure

Information and Communication Technology

Areas of Advance

Transport

Areas of Advance

Health Engineering

Areas of Advance

Publications

More information

Latest update

8/19/2021