Umeå University is one of Sweden’s largest institutions of higher education with over 37,000 students and 4,800 faculty and staff. We are characterised by world-leading research in several scientific fields and a multitude of educations ranked highly in international comparison. Umeå University is also the site of the pioneering discovery of the CRISPR-Cas9 genetic scissors - a revolution in genetic engineering that has been awarded the Nobel Prize in Chemistry.
At Umeå University, everything is nearby. Our cohesive campus environment makes it easy to meet, collaborate and exchange knowledge, which promotes a dynamic and open culture where we rejoice in each other's successes.
Are you interested in knowing more about Umeå University as a workplace read more at: Work with us.
The Department of Radiation Sciences is, in terms of research, a dynamic and internationally successful environment in radiology, oncology, cognitive neuro-science, radiation physics and bio-medical engineering. The department belongs to the Faculty of Medicine at Umeå University.
The Department of Radiation Sciences welcome applications for postdoctoral studies in Artificial Intelligence in healthcare. The position is for two years full-time, starting by agreement.
The project
Advancements in artificial intelligence (AI) have the possibility to significantly contribute to improve patient treatments, in a trend towards personalized and predictive medicine. Indeed, while deep learning has shown its potential in many fields, healthcare included, we have assisted to extensive research of new quantitative biomarkers computed from the medical data already collected in clinical practice. However, most of the AI models consider only unimodal data, while there is the need to investigate the fusion between the heterogenous nature of healthcare data, such as images, molecular biomarkers, clinical data, and electronic health records (EHRs). Within this project, we are interested in studying multimodal deep learning as well as in explaining the decisions taken by this paradigm, two areas of great interest at its infancy. In particular, in multimodal deep learning we aim to address open scientific questions, i.e., which modalities are useful, how to fuse them and where this integration should happen, how to embed in the training any process able to learn more powerful data representations, how to get the optimal fusion architecture robust to missing data and missing modalities. As for multimodal explanations, there is the need to help physicians, regulators, and patients to trust AI models. However, the few efforts in multimodal explanations available in the literature have been directed towards computer vision and natural language applications, so that multimodal explanations are completely missing in healthcare. This project therefore addresses this challenge, particularly focusing on pre-modeling, during modelling and post-hoc modeling explanations at local level.
These topics will be applied to true data collected in the context of precision oncology, e.g., in lung cancer, where the prognosis variability, due to inter-patient, inter-tumor and intra-tumor heterogeneity, is a challenge for the scientific community. Hence, better prognostic and predictive tools to tailor treatment decisions for cancer patients are needed in order to maximize outcome and minimize unnecessary side-effects. To this end, multimodal deep learning and multimodal explanations will help us discovering new quantitative biomarkers from the heterogeneous digital patient phenotypes that are routinely collected in the clinical practice, and that we have already collected in previous research projects.
Work responsibilities
The researcher will design and implement multimodal learning algorithms, design and implement methods to provide explanations of the decisions taken, but he/she can also adapt and improve existing algorithms and methods. The Postdoctor will oversee data collection carried out the physicians, maintain the data and check their quality, prepare and analyse research data,summarize experiments, use graphics and any useful software to analyse and present data, prepare scientific manuscript for submission to international conferences and peer-reviewed journals, participate in project meetings and conferences. Working in vibrant and exciting environment the Postdoctor will also co-supervise PhD and master’s students.
Eligibility
A person who has been awarded a doctorate or a foreign qualification deemed to be the equivalent of a doctorate in one of these fields: artificial intelligence, machine learning, computer science, biomedical engineering is eligible for appointment as postdoctoral researcher. This eligibility requirement must be met no later than the time at which the appointment decision is made. The applicant's degree, or through supplementary education, should have a specialization that includes,or combines, computer science with an AI or machine learning focus. The ideal candidate has a background in deep learning and machine learning, has experience with 1D, 2D, 3D medical data, such as images, as well as with tabular data, is able to write and/or critically review a scientific manuscript with high degree of autonomy.
Previous experience with multimodal learning as well as with eXplainable AI techniques is very well welcome.
Other qualifications
Since appointment as a postdoctoral researcher is a career-development position for junior researchers, we are primarily interested in applicants who completed their doctoral degree no more than three years before the application deadline. The following qualifications are recommended:
- MSc and PhD degree in the field of applied mathematics, computer science/informatics, (biomedical) engineering, (medical) physics or comparable applied / natural sciences
- Knowledge in machine learning and deep learning, mostly for healthcare data Strong programming skills especially in python and the relevant deep learning libraries (pytorch or tensorflow) Experience in git and other version control
- Good publications track record in the relevant journals (e.g., Pattern Recognition, Medical Image Analysis, Information Fusion, Artificial Intelligence in Medicine, IEEE Transactions in Medical Imaging, Knowledge-Based Systems, etc.)
- Experience and willingness of working in a multidisciplinary team, incl. clinicians and medical physicists
- Strong scientific communication and presentation skills in English
Application
The application must include:
- A brief description of your research interests and a statement describing why you are interested in the position.
- CV
- List of publications
- Degree certificate from doctoral studies and other relevant degrees
- Contact details of two reference persons
The application must be written in English. The application must be made via our e-recruitment latest November 18, 2022.
We look forward to receiving your application!
Information box
Admission
By agreement
Salary
Monthly
Apply before
2022-11-18
Registration number
AN 2.2.1-1586-22
Contact
prof. Paolo Soda
paolo.soda@umu.se
Union representative
SACO
090-7865365
SEKO
090-7865296
ST
090-7865431
No comments:
Post a Comment