Postdoctoral Fellowship (05) at Umea University, Sweden

The project investigates the long-term consequences of COVID-19, known as post-COVID condition (PCC). The aim is to understand why some people develop long-lasting sequalae and how these the prognosis vary between phenotypes. Our goals are to:

1. Identify different types of PCC in order to provide more tailored rehabilitation and treatment.
2. Develop methods and models that can predict the course of the disease by analyzing detailed data on the immune system and metabolism.

We are conducting a large study, the CoVUm study, involving 579 patients with both mild and severe COVID-19 infection. They are being followed for three years after the infection.

This project proposes the implementation of a cognitive agent embodied in a humanoid robot which will collaborate with a human partner or with another humanoid robot to solve a spatial problem (e.g. 3D puzzle, fold a paper). The tasks to be carried out are: (i) scene understanding: detection of objects and relations between objects, and use of these relations to infer new knowledge (i.e. reasoning); (ii) explore object affordances, learn the consequences of the actions carried out and enrich the knowledge base (i.e. learning by interaction); (iii) querying the knowledge base about what was useful in the past to predict actions that might be useful in the present, try them out and update its knowledge base (i.e. update learning). Also writing scientific papers and communicating our research advances in conferences. Methods: programming a humanoid platform using ROS2 packages, solve SLAM, use imitation learning algorithms to learn pick-and-place actions, design HRI experiments with users, evaluate data, and share the code and benchmarks in open repositories.

This position is funded by the European project COMFORT (www.comfort-ai.eu) and the AI Policy Lab in Umeå University (www.aipolicylab.se). COMFORT (Improve Diagnosis and Treatment of Urologic Cancers with Artificial Intelligence-Driven Decision Support Tool) is developing a cutting-edge decision support system using artificial intelligence (AI) and data-driven approaches to assist medical professionals in delivering improved care for people affected by prostate cancer (PCa) or kidney cancer (KC). 

The first phase of this postdoctoral research will focus on modeling AI trustworthiness and explainability in medical data and models, with particular emphasis on fairness and privacy. The second phase will explore integrating AI governance and policy with the outcomes achieved in the first phase.

The Department of Community Medicine and Rehabilitation is seeking a Postdoc for the project “COPD-HIT”, an international multicenter project aimed at developing and validating an in vitro loading model against in vivo blood and muscle adaptations (biopsies) from individuals with Chronic Obstructive Pulmonary Disease (COPD), as well as healthy matched controls who have undergone physical training of varying intensity and duration.

The project will develop privacy-aware machine learning (ML) models. We focus on data-driven models for complex and temporal data, including those built from synthetic sources. Within privacy, we are interested in different types of privacy measures and models (differential and integral privacy, k-anonymity), different scenarios (centralized and decentralized data; local and global privacy). For decentralized data, we consider federated learning. We are interested in privacy-preserving machine learning at a scale.