Dynamic data modelling

The research project Data-driven identification of dynamical systems, with applications in Space Physics, aims to combine classical and numerical methods from the theory of differential equations and inverse problems with new emerging tools in computational data science including Approximative Bayesian Computation to obtain interpretable and informative learning systems capable of predicting complex physical processes. The research will concentrate on developing and applying methodologies of data science and artificial intelligence to tackle research problems related to space climate and space weather, i.e., disturbances of the near-Earth space and ionosphere driven by the solar wind from the Sun. Topics include dynamics of geospace current systems and energetic particle populations based on ground-based and satellite observations.

The study group features close collaboration with the Space Climate and Ionospheric Physics groups, and the Finnish Centre of Excellence in Inverse Modelling and Imaging.

Postdoctoral researcher: Dr. Snizhana Ross

Study group members:

Prof. Anita Aikio (Space Physics)
Assoc. Prof. Timo Asikainen (Space Physics)
Assoc Prof. Andreas Hauptmann (Mathematical Sciences)
Prof. Mikko Sillanpää (Mathematical Sciences)
Dr. Ilkka Virtanen (Space Physics)
Dr. Heikki Vanhamäki (Space Physics)

The research project data driven prediction and interpretation of spectra, applications in climate research, aims to develop novel data analysis tools for analysing spectroscopic and spectromicroscopic data acquired at large scale research facilities, such as the MAX IV Laboratory. More specifically, the focus is on developing new methods of extracting element and compound specific, as well as chemical functional group information about the samples from the spectroscopic and spectromicroscopic data sets. The samples can range from both biological and inorganic samples from various industries to atmospheric aerosols and nanoparticles.

The study group features close collaboration with the Nano and Molecular Systems Research Unit (NANOMO).

Postdoctoral researcher:

Dr. Mohammed Alaoui Mansouri (NANOMO)

Study group members:

Dr. Samuli Urpelainen (NANOMO)
Dr. Mourad Kharbach (Neste Corporation, Finland)
Dr. Evgeniy Redekop (Deparment of Chemistry, University of Oslo)
Prof. Mikko Sillanpää (Mathematical Sciences)
Dr. Ilkka Launonen (Mathematical Sciences)
M.Sc. Olli Sarala (Mathematical Sciences)

The research project “quantitative genetics: genomic predictions in forest tree breeding” aims to combine expertise in quantitative genetic methodology and forest genomics and breeding. The research will concentrate on developing and applying new methodologies of statistical learning to tackle research problems related to genomic predictions in forest tree breeding. An on-going investigation focuses on the genetic associations with chemical phenotypes and genomic selection of Pinus sylvestris (the Scots pine).

The study group features close collaboration with the Research Unit of Ecology and Genetics and Natural Resource Institute of Finland (LUKE), and Department of Forest Sciences (University of Helsinki).

Postdoctoral researcher:

Dr. Kalle Leppälä (Organismal and Evolutionary Biology, University of Helsinki)

Study group members:

Prof. Juha Heikkinen (LUKE)
Dr. Chedly Kastally (Forest Sciences, University of Helsinki)
Dr. Sonja Kujala (Ecology and Genetics)
Prof. Katri Kärkkäinen (LUKE)
Dr. Alina Niskanen (Ecology and Genetics)
Assoc. prof. Tanja Pyhäjärvi (Forest Sciences, University of Helsinki)
Prof. Mikko Sillanpää (Mathematical Sciences)

The research project “identification of the genetic predispositions of physical and mental (longitudinal) co-morbidities” aims to combine methods of quantitative genetics and life-course epidemiology and apply genetically informed methods to elucidate the comorbidity observed between mental and cardiometabolic traits in humans.

The position is shared between the Research Unit of Mathematical Sciences and the Center for Life Course Health Research (ELITE) and partly funded by The H2020-EarlyCause project – A research and innovation project funded by the European commission to analyse the causative factors pertaining the early origins of co-morbidities between mental (e.g. depression, anxiety, schizophrenia) and cardio-metabolic (e.g. type 2 diabetes, atherosclerosis, obesity) diseases.

Postdoctoral researcher: Dr. Ville Karhunen

The research project "interpretable artificial intelligence with applications in medical imaging and knee arthritis prediction" aims to develop methods in pattern recognition and machine learning for early detection of knee arthritis from medical imaging data. The focus of the study group is to develop and employ interpretable artificial intelligence techniques which can cast light on the early-onset signs of knee arthritis.

The position is jointly funded by Profi5/HiDyn and Profi5/DigiHealth. The study group features close collaboration with the research unit of Medical Imaging, Physics and Technology at the Faculty of Medicine, University of Oulu.

Postdoctoral researcher: Dr. Juha Tiirola

Artificial intelligence can be used for predicting very complex and high dimensional phenomena. However, due to the complexity of AI methods, it is often unclear why AI proposed a given prediction with a given set of predictors and how much uncertainty there is in the prediction. Such a lack of transparency and interpretability hampers the usefulness of AI in many applications. As an example, in medical imaging it is important for the doctor to be able to evaluate the uncertainty in the predictions. On the other hand, in the industry, changes in process lines cost money and therefore it is important to understand why a given change is suggested. For these reasons, the research project "enhancing transparency and understandability of artificial intelligence methods" aims to develop methods that facilitate measuring the uncertainty in the predictions and enhancing the understandability and interpretability of AI methods.

The study group features close collaboration with the Center for Machine Vision and Signal Analysis (CMVS), the Center for Ubiquitous Computing (UBICOMP) and the Biomimetics and Intelligent Systems Group (BISG) at the Faculty of Information Technology and Electrical Engineering, University of Oulu.

Postdoctoral researcher: Tero Lähderanta

The research project "data driven dynamic modelling of microstructures in steel” aims at developing novel data driven methods for dynamic modelling of microstructures in steel.

The study group features close collaboration with the Materials and Mechanical Engineering Unit and Process Metallurgy research unit at the University of Oulu.

Doctoral researcher: M.Sc. Juho Luukkonen

Study group members:

Dr. Erkki Laitinen (Mathematical Sciences)
Prof. Seppo Louhenkilpi (Process Metallurgy)
Dr. Aarne Pohjonen (Process Metallurgy)
Prof. Mikko Sillanpää (Mathematical Sciences)

The research project "developing methods for complex multipoint calibration, applications in radio frequency (RF) system architectures" aims at developing and applying methodologies of data science and artificial intelligence to tackle research problems related to multipoint calibration problems. More specifically, the aim is to optimize system performance in extremely complex large-scale mm-wave RF architectures where high-dimensional control parameters typically operate non-linearly, time varying and can have high dimensional interactions in noisy environment (that is, network of interdependencies is extremely complex). At a large, the problem can be seen also as an online control problem where control data is accumulating dynamically.

The study group features close collaboration with the Centre for Wireless Communications (CWC) and Center for Machine Vision and Signal Analysis (CMVS) at the Faculty of Information Technology and Electrical Engineering, University of Oulu.

Postdoctoral researcher: Dr. Krishna Chaitanya Bulusu (CWC)

Study group members:

Prof. Markku Juntti (CWC)
Dr. Ilkka Launonen (Mathematical Sciences)
Prof. Aarno Pärssinen (CWC)
Prof. Mikko Sillanpää (Mathematical Sciences)
Prof. Olli Silvén (CMVS)
Dr. Visa Tapio (CWC)
Dr. Nuutti Tervo (CWC)

• Associate professor: Dr. Patrik Waldmann

Dr. Patrik Waldmann works with applied computational statistics. His research involves high dimensional data analysis, statistical machine learning, Bayesian statistics, statistical genetics and bioinformatics, and precision modelling in agriculture, medicine and technology.

• Doctoral student in high-dimensional causal inference with applications in genomics: M.Sc. Zewude Berkessa
• Postdoctoral researcher in multi-task learning in bioinformatics:
  Dr. Yuhua Fan
• Postdoctoral researcher in deep learning for genomic predictions: Dr. Ragini Gokhale

• Adjunct professor: Dr. Xiaobai Li

Adjunct Professor Xiaobai Li works on multimodal emotion analysis. Her research focuses on high-dimensional data fusion for human emotion understanding by analyzing all visual dynamic changes. Related sub-topics include, e.g., facial expression/micro-expression analysis, body movements recognition, and remote physiological signal measurement from facial videos.

• Postdoctoral researcher in measuring physiological signals from facial expression videos: Puneet Kumar

• Associate professor: Dr. Miguel Bordallo Lopez

Associate Professor Miguel Bordallo work in the field of Vision Systems Engineering. His current research focuses on using real-time computer vision and multimodal sensing in multiple application domains, with a special interest in digital health. Related sub-topics include radio-based sensing, thermal and infrared imaging, embedded and distributed computing, and multimodal data fusion.

• Doctoral or postdoctoral researcher: position open

Chair:

Prof. Mikko J. Sillanpää, Research Unit of Mathematical Sciences

Vice chair:

Prof. Olli Silvén, Center for Machine Vision and Signal Analysis

Dr. Ilkka Launonen, Research Unit of Mathematical Sciences

Prof. Saija Luukkanen, Oulu Mining School

Assoc. Prof. Andreas Hauptmann, Research Unit of Mathematical Sciences

Dr. Minna Ruddock, Northern Finland Birth Cohort Studies

Prof. Olli Silvén, Center for Machine Vision and Signal Analysis