Embedded Files
If you are enrolled in a master's program and interested in writing your thesis in collaboration with the UiT Machine Learning Group, we propose some interesting projects you can get involved with. You may also contact directly one of the group members if you wish to work on a topic that is not listed here.
Theoretical Machine Learning and an application to marine science
Theoretical Machine Learning and an application to marine science
2 Master projects in collaboration with the University of Oslo:
Formulating Explainability Principles for a Self-Explainable Model Through Information Theoretic Learning
Formulating Explainability Principles for a Self-Explainable Model Through Information Theoretic Learning
Self-supervised Multi-resolution Partitioning of Marine Acoustics Semantics
Self-supervised Multi-resolution Partitioning of Marine Acoustics Semantics
Using Deep Learning based Iris Recognition for Wildlife Monitoring
Using Deep Learning based Iris Recognition for Wildlife Monitoring
Iris recognition is a method of biometric identification that uses information from the irises of an individual's eyes to extract complex patterns that are unique and stable. Iris recognition is a mature field when it comes to recognition of humans, however, its use for wildlife animal and bird monitoring is still evolving.
In this project, the aim is to adapt the iris recognition methodology for recognizing sea birds that are otherwise identified by using a numbered metal or plastic tag. For this, we need to use a deep learning-based pipeline with modifications suitable for the task. A small dataset with Svalbard birds has been made available that can be used for the development of this deep learning pipeline.
Contact Person
Puneet Sharma (puneet.sharma@uit.no)
Deep learning aided quantification in PET imaging
Deep learning aided quantification in PET imaging
Positron emission tomography (PET) is a medical imaging technique that visualises the distribution of an injected radioactive tracer in living subjects. Medical imaging with PET plays an important role in the detection, staging, and treatment response assessment of many diseases, including cancer, neurological and cardiovascular conditions, as well as inflammation and infection.
PET is quantitative, in the sense that it allows not only visualisation but also non-invasive quantification of regional tracer uptake. Specifically, with dynamic PET imaging, it is possible to fully assess the time-dependent tracer distribution in the body. This allows quantification of biological processes, such as glucose metabolism or blood flow, by using tracer kinetic modelling. Tracer kinetic modelling, however, requires accurate determination of an arterial input-function (AIF), i.e. the tracer time-activity curve in blood.
The gold-standard AIF is obtained by measuring the time-dependent FDG radioactivity concentration in arterial blood through invasive blood sampling. This procedure is complex, time-consuming, and potentially painful with risk for complications.
The aim of this Master´s project is to develop novel methodology for arterial input function prediction, building upon state-of-the-art deep learning methods. In addition, the project will focus on developing interpretability and uncertainty methods to explain outcomes and the variability in the predictions. Both preclinical (mice) and clinical (human) PET data will be explored in the project.
Prerequisites:
Relevant machine learning courses, for instance FYS-2021 and FYS-3033
Programming skills in Python, preferably using PyTorch and/or TensorFlow
Experience with medical image processing is an advantage
Contact persons
Samuel Kuttner (samuel.kuttner@uit.no)
Graph neural networks
Graph neural networks
We have several research projects focused on graphs and graph machine learning. The main topics are:
theoretical concepts, connecting Math and Physics to Machine Learning,
using graph neural net for generative AI,
application of graph neural nets to weather modelling,
application to document analysis.
See our Graph ML Group website for more details:
https://ngmlgroup.github.io/theses.html
Recommended prerequisites: Knowledge of machine learning, deep learning and signal/image processing. Good programming skills (Python).
Contact person: Benjamin Ricaud (benjamin.ricaud@uit.no)
Population counting using Drone Images for Marine Surveys
Population counting using Drone Images for Marine Surveys
Marine surveys require use of valuable resources (expert's time and boats). UiT in collaboration with Norwegian Polar Institute and University of Southern Denmark is working towards developing a solution for performing population counting based on images captured from flying a drone. The initial plan is to develop a supervised learning based methodology for detecting the number of porpoises in an image. Later on, the plan is to further develop the framework to accommodate for other similar mammal species (with fewer training samples).
Prerequisites: FYS-2021, FYS-3033
Contact person: Puneet Sharma (puneet.sharma@uit.no)
Precise Uncertainty Quantification in Dynamic Line Rating Forecasts
Precise Uncertainty Quantification in Dynamic Line Rating Forecasts
Description
In this master's thesis, the student will focus on developing a model to forecast line temperature, a crucial factor in Dynamic Line Rating (DLR) for energy grids. The model will utilize sensor data (specifically line current) and weather data, including both historical and forecasted information sourced from YR APIs. The data will be pre-processed and provided to the student. The key challenge is to not only accurately forecast line temperatures but also to precisely quantify the associated uncertainties.
Objective
The main objective is to create a model that accurately forecasts line temperature by using sensor and weather data. Emphasis will be placed on quantifying and controlling the forecast's uncertainty.
Approach
Model Development: Apply machine learning techniques with a particular focus in uncertainty quantification.
Uncertainty Analysis: Develop methods to rigorously quantify the uncertainty bounds in line temperature forecasts and their subsequent influence on DLR predictions.
Potential Impact
This project will significantly enhance DLR in energy grid management by providing more accurate and reliable line temperature forecasts. Such advancements will facilitate better decision-making, improving the efficiency and dependability of power distribution networks.
Prerequisites
A solid understanding of machine learning, deep learning, and statistical analysis. Proficiency in programming, ideally in Python. Background knowledge in energy systems or meteorological data analysis is an advantage but not essential.
Contact Person
Puneet Sharma (puneet.sharma@uit.no)
Industry partner: Avju Solutions, Jonas Mørch-Lampe (jml@avju.com)
On the impact of the missigness bias in masking-based explainability methods
On the impact of the missigness bias in masking-based explainability methods
Description
Masking-based methods is a popular approach approach for explaining the decision of deep learning algorithms. The key idea is to mask out parts of the input and monitor how the prediction of the model changes. However, removing input parts have been shown to introduce biases in convolutional neural networks, which could lead to artefacts in the explanations. This project will explore the implications of the missingess bias in masking-based methods for explainability and investigate ways to address this bias.
Contact Person
Kristoffer Wickstrøm (kristoffer.k.wickstrom@uit.no)
Google Sites
Report abuse