Machine learning & AI talks
The UiT Machine Learning Group is continuously organizing talks with the aim of bringing together researchers across departments as well as external stakeholders with an interest in machine learning and AI. The talks are intended to span broadly over mathematical methodology development, scalability, and various applications.
Every Wednesday the Machine learning group at UiT hosts a lunch meeting where a speaker gives a talk on a topic in machine learning or AI research. Usually the speakers are members of the group that present some of their recent work, updates or an interesting research paper they want to share. The Wednesday meetings also hosts external speakers that are invited to give a presentation. On this page we present highlights of some recent talks.
If you are interested in presenting your work, please contact Jonas Nordhaug Myhre.
Estimation of F-region plasma density and virtual height of the F-region in ionograms using deep learning
Speaker: Juha Vierinen, UiT Space Physics Group4th February 2021
Juha Vierinen, associate professor at the UiT Space Physics Group, gave a lunch talk on deep learning-based analysis of ionograms. The goal is to automatically estimate the plasma density and virtual height of the F-region, which would enable large scale statistical analysis of ionogram imagery.
A study of generative adversarial networks to improve classification of microscopic foraminifera
Speaker: Gitta Kutynoik, Ludwig-Maximilians-Universität München27th January 2021
Gitta Kutyniok, Bavarian AI Chair for Mathematical Foundations of Artificial Intelligence, LMU Munich, gave a lunch talk on recent research on both graph neural networks and explainability in deep learning. The talk can be viewed here.
A study of generative adversarial networks to improve classification of microscopic foraminifera
Speaker: Eirik Agnalt Østmo16th September 2020
Foraminifera are single-celled organisms with shells that live in the marine environment and can be found abundantly as fossils in e.g. sediment cores. The assemblages of different species and their numbers serves as an important source of data for marine, geological, climate and environmental research.
In this week's internal talk Eirik presents the work from his master's thesis where he attempted to use generative adversarial networks (GANs) to model microscope images of foraminifera and generate synthetic foraminifera images. The synthetic images was used in addition to the original images to attempt to improve a deep learning-based classification model that had previously been proposed (Johansen and Sørensen, 2020).
Interpolations in the latent space of a GAN trained on planktic foraminifera.
Towards a Framework for Noctilucent Cloud Analysis
Speaker: Puneet Sharma2nd September 2020
Noctilucent clouds (NLC) that can be seen by the naked eye and polar mesospheric summer echoes (PMSE) observed with radar display the complex dynamics of the atmosphere at approximately 80 to 90 km height. Both the NLC and PMSE phenomena occur in the presence of ice particles. The topic of ice particle formation and its link to climate change is currently in debate in the atmospheric community and needs further investigation. The morphological shape of the clouds indicates the influence of wave propagation of various scales. Proper description and comparison of cloud structure is a step toward a detailed analysis of the role of atmospheric waves in the formation of PMSE and NLC. As a first step towards an integrated statistical analysis, we propose a framework for the analysis of NLC observation.
Noctilucent cloud activity correctly predicted in the red squares by the proposed CNN algorithm.
The Approximation Power of Deep Neural Networks: Theory and Applications
Speaker: Gitta Kutynoik, Technische Universität Berlin16th August 2019
Despite the outstanding success of deep neural networks in real-world applications, most of the related research is empirically driven and a mathematical foundation is almost completely missing. The main goal of a neural network is to approximate a function, which for instance encodes a classification task. Thus, one theoretical approach to derive a fundamental understanding of deep neural networks focusses on their approximation abilities.
In this talk we will provide an introduction into this research area. After a general overview of mathematics of deep neural networks, we will discuss theoretical results which prove that not only do (memory-optimal) neural networks have as much approximation power as classical systems such as wavelets or shearlets, but they are also able to beat the curse of dimensionality. On the numerical side, we will then show that superior performance can typically be achieved by combining deep neural networks with classical approaches from approximation theory.
Learning the Invisible: Taking the Best out of the Data and Modeling World
Speaker: Gitta Kutynoik, Technische Universität Berlin15th August 2019
Pure model-based approaches are today often insufficient for solving complex inverse problems in imaging. At the same time, we witness the tremendous success of data based methodologies, in particular, deep neural networks for such problems. However, at the same time, pure deep learning approaches often neglect known and valuable information from the modeling world.
In this talk, we will provide an introduction into this problem complex and then focus on the inverse problem of (limited-angle) computed tomography. We will develop a conceptual approach by combining the model-based method of sparse regularization by shearlets with the data-driven method of deep learning. Our solvers are guided by a microlocal analysis viewpoint to pay particular attention to the singularity structures of the data. Finally, we will show that our algorithm significantly outperforms previous methodologies, including methods entirely based on deep learning.
Low-Dose CT Grand Challenge data set.
Applied Harmonic Analysis meets Sparse Regularization of Inverse Problems
Speaker: Gitta Kutynoik, Technische Universität Berlin13th August 2019
Sparse regularization of inverse problems has already shown its effectiveness both theoretically and practically. The area of applied harmonic analysis offers a variety of systems such as wavelet systems which provide sparse approximations within certain model situations which then allows to apply this general approach provided that the solution belongs to this model class. However, many important problem classes in the multivariate situation are governed by anisotropic structures such as singularities concentrated on lower dimensional embedded manifolds, for instance, edges in images or shear layers in solutions of transport dominated equations. Since it was shown that the (isotropic) wavelet systems are not capable of sparsely approximating such anisotropic features, the need arose to introduce appropriate anisotropic representation systems. Among various suggestions, shearlets are the most widely used today. Main reasons for this are their optimal sparse approximation properties within a suitable model situation in combination with their unified treatment of the continuum and digital realm, leading to faithful implementations.
In this talk, we will first provide an introduction to sparse regularization of inverse problems, followed by an introduction to the area of applied harmonic analysis, in particular, discussing the anisotropic representation system of shearlets and presenting the main theoretical results. We will then analyze the effectiveness of using shearlets for sparse regularization of exemplary inverse problems both theoretically and numerically.
Ocean exploration with machine learning: An Antidote to Chaos?
Speaker: Maike Sonnewald26th February 2019
Machine learning has the potential to revolutionize oceanography, if applied with care. Three case studies highlight the potential for greatly accelerating the efficiency of ocean exploration using supervised (neural networks) and unsupervised (clustering) machine learning. Firstly, two decades of data from the realistic ECCO state estimate 3D physical fields are used to objectively determine global physical regimes using k-means clustering. The identified regions corresponds closely to those predicted by canonical theory from physical oceanography and the method can be scaled to analyze vast amounts of data from e.g. the Climate Model Intercomparison Project. Secondly, the highdimentional dataset from the biogoechemical DARWIN model reveals the existence of ecological niches using t-SNE and DBSCAN clustering. Constraining ocean biomes, niches and larger regions can be examined to understand how sensitive they are to climate forcing which is crucial to protecting our ocean.
Speaker: Henrik Boström14th February 2019
Conformal prediction is a relatively new framework in which the predictive models output sets of predictions with a bound on the error rate, i.e., in a classification context, the probability of excluding the correct class label is lower than a predefined significance level. An investigation of the use of decision trees within the conformal prediction framework is presented, with the overall purpose to determine the effect of different algorithmic choices, including split criterion, pruning scheme and way to calculate the probability estimates. Since the error rate is bounded by the framework, the most important property of conformal predictors is efficiency, which concerns minimizing the number of elements in the output prediction sets. Results from one of the largest empirical investigations to date within the conformal prediction framework are presented, showing that in order to optimize efficiency, the decision trees should be induced using no pruning and with smoothed probability estimates. The choice of split criterion to use for the actual induction of the trees did not turn out to have any major impact on the efficiency. Finally, the experimentation also showed that when using decision trees, standard inductive conformal prediction was as efficient as the recently suggested method cross-conformal prediction. This is an encouraging results since cross-conformal prediction uses several decision trees, thus sacrificing the interpretability of a single decision tree.