Umeå University's logo

umu.sePublications
Change search
Link to record
Permanent link

Direct link
Alternative names
Publications (10 of 58) Show all publications
Meyers, C., Saleh Sedghpour, M. R., Löfstedt, T. & Elmroth, E. (2025). A training rate and survival heuristic for inference and robustness evaluation (Trashfire). In: Proceedings of 2024 International Conference on Machine Learning and Cybernetics: . Paper presented at 2024 International Conference on Machine Learning and Cybernetics (ICMLC),Miyazaki, Japan, September 20-23, (pp. 613-623). IEEE
Open this publication in new window or tab >>A training rate and survival heuristic for inference and robustness evaluation (Trashfire)
2025 (English)In: Proceedings of 2024 International Conference on Machine Learning and Cybernetics, IEEE, 2025, p. 613-623Conference paper, Published paper (Refereed)
Abstract [en]

Machine learning models—deep neural networks in particular—have performed remarkably well on benchmark datasets across a wide variety of domains. However, the ease of finding adversarial counter-examples remains a persistent problem when training times are measured in hours or days and the time needed to find a successful adversarial counter-example is measured in seconds. Much work has gone into generating and defending against these adversarial counter-examples, however the relative costs of attacks and defences are rarely discussed. Additionally, machine learning research is almost entirely guided by test/train metrics, but these would require billions of samples to meet industry standards. The present work addresses the problem of understanding and predicting how particular model hyper-parameters influence the performance of a model in the presence of an adversary. The proposed approach uses survival models, worst-case examples, and a cost-aware analysis to precisely and accurately reject a particular model change during routine model training procedures rather than relying on real-world deployment, expensive formal verification methods, or accurate simulations of very complicated systems (e.g., digitally recreating every part of a car or a plane). Through an evaluation of many pre-processing techniques, adversarial counter-examples, and neural network configurations, the conclusion is that deeper models do offer marginal gains in survival times compared to more shallow counterparts. However, we show that those gains are driven more by the model inference time than inherent robustness properties. Using the proposed methodology, we show that ResNet is hopelessly insecure against even the simplest of white box attacks.

Place, publisher, year, edition, pages
IEEE, 2025
Series
Proceedings (International Conference on Machine Learning and Cybernetics), ISSN 2160-133X, E-ISSN 2160-1348
Keywords
Machine Learning, Computer Vision, Neural Networks, Adversarial AI, Trustworthy AI
National Category
Artificial Intelligence Security, Privacy and Cryptography Computer Sciences
Identifiers
urn:nbn:se:umu:diva-237109 (URN)10.1109/ICMLC63072.2024.10935101 (DOI)2-s2.0-105002274020 (Scopus ID)9798331528041 (ISBN)9798331528058 (ISBN)
Conference
2024 International Conference on Machine Learning and Cybernetics (ICMLC),Miyazaki, Japan, September 20-23,
Funder
Knut and Alice Wallenberg Foundation, 2019.0352eSSENCE - An eScience Collaboration
Available from: 2025-04-02 Created: 2025-04-02 Last updated: 2025-05-19Bibliographically approved
Lindström, M., Blöcker, C., Löfstedt, T. & Rosvall, M. (2025). Compressing regularized dynamics improves link prediction with the map equation in sparse networks. Physical review. E, 111(5), Article ID 054314.
Open this publication in new window or tab >>Compressing regularized dynamics improves link prediction with the map equation in sparse networks
2025 (English)In: Physical review. E, ISSN 2470-0045, E-ISSN 2470-0053, Vol. 111, no 5, article id 054314Article in journal (Refereed) Published
Abstract [en]

Predicting future interactions or novel links in networks is an indispensable tool across diverse domains, including genetic research, online social networks, and recommendation systems. Among the numerous techniques developed for link prediction, those leveraging the networks' community structure have proven highly effective. For example, the recently proposed MapSim predicts links based on a similarity measure derived from the code structure of the map equation, a community-detection objective function that operates on network flows. However, the standard map equation assumes complete observations and typically identifies many small modules in networks where the nodes connect through only a few links. This aspect can degrade MapSim's performance on sparse networks. To overcome this limitation, we propose to incorporate a global regularization method based on a Bayesian estimate of the transition rates along with three local regularization methods. The regularized versions of the map equation compensate for incomplete observations and mitigate spurious community fragmentation in sparse networks. The regularized methods outperform standard MapSim and several state-of-the-art embedding methods in highly sparse networks. This performance holds across multiple real-world networks with randomly removed links, simulating incomplete observations. Among the proposed regularization methods, the global approach provides the most reliable community detection and the highest link prediction performance across different network densities. The principled method requires no hyperparameter tuning and runs at least an order of magnitude faster than the embedding methods.

Place, publisher, year, edition, pages
American Physical Society, 2025
National Category
Statistical physics and complex systems Other Computer and Information Science
Identifiers
urn:nbn:se:umu:diva-239088 (URN)10.1103/physreve.111.054314 (DOI)2-s2.0-105005834751 (Scopus ID)
Funder
Swedish Research Council, 2022-06725Swedish Research Council, 2023-03705Knut and Alice Wallenberg Foundation
Available from: 2025-05-23 Created: 2025-05-23 Last updated: 2025-06-02Bibliographically approved
Ghanbari Azar, S., Tronchin, L., Simkó, A., Nyholm, T. & Löfstedt, T. (2025). From promise to practice: a study of common pitfalls behind the generalization gap in machine learning. Transactions on Machine Learning Research
Open this publication in new window or tab >>From promise to practice: a study of common pitfalls behind the generalization gap in machine learning
Show others...
2025 (English)In: Transactions on Machine Learning Research, E-ISSN 2835-8856Article in journal (Refereed) Published
Abstract [en]

The world of Machine Learning (ML) offers great promise, but often there is a noticeable gap between claims made in research papers and the model's practical performance in real-life applications. This gap can often be attributed to systematic errors and pitfalls that occur during the development phase of ML models. This study aims to systematically identify these errors. For this, we break down the ML process into four main stages: data handling, model design, model evaluation, and reporting. Across these stages, we have identified fourteen common pitfalls based on a comprehensive review of around 60 papers discussing either broad challenges or specific pitfalls within ML pipeline. Moreover, Using the Brain Tumor Segmentation (BraTS) dataset, we perform three experiments to illustrate the impacts of these pitfalls, providing examples of how they can skew results and affect outcomes. In addition, we also perform a review to study the frequency of unclear reporting regarding these pitfalls in ML research. The goal of this review was to assess whether authors have adequately addressed these pitfalls in their reports. For this, we review 126 randomly chosen papers on image segmentation from the ICCV (2013-2021) and MICCAI (2013-2022) conferences from the last ten years. The results from this review show a notable oversight of these issues, with many of the papers lacking clarity on how the pitfalls are handled. This highlights an important gap in current reporting practices within the ML community. The code for the experiments is available at https://github.com/SG-Azar/BraTS-ML-Pitfalls-Experiments.

Place, publisher, year, edition, pages
Transactions on Machine Learning Research, 2025
National Category
Computer and Information Sciences
Identifiers
urn:nbn:se:umu:diva-233898 (URN)2-s2.0-85219528926 (Scopus ID)
Funder
Swedish Childhood Cancer Foundation, MT2021-0012Lions Cancerforskningsfond i Norr, LP 22-2319Lions Cancerforskningsfond i Norr, LP 24-2367
Available from: 2025-01-10 Created: 2025-01-10 Last updated: 2025-03-18Bibliographically approved
Löfstedt, T. (2025). Matematikångest och bristande förkunskaper. In: Agneta Bränberg; Dan Borglund (Ed.), Didaktisk dialog: i kursen Didaktik för universitetslärare (pp. 112-117). Umeå: Umeå University
Open this publication in new window or tab >>Matematikångest och bristande förkunskaper
2025 (Swedish)In: Didaktisk dialog: i kursen Didaktik för universitetslärare / [ed] Agneta Bränberg; Dan Borglund, Umeå: Umeå University, 2025, p. 112-117Chapter in book (Other academic)
Place, publisher, year, edition, pages
Umeå: Umeå University, 2025
Series
Skriftserie från Universitetspedagogik och lärandestöd (UPL) ; 2025:1
National Category
Didactics
Identifiers
urn:nbn:se:umu:diva-234732 (URN)978-91-8070-614-8 (ISBN)
Available from: 2025-01-29 Created: 2025-01-29 Last updated: 2025-01-29Bibliographically approved
Lin, D., Hägg, L., Wadbro, E., Berggren, M. & Löfstedt, T. (2025). Structured regularization using approximate morphology for Alzheimer's disease classification. In: 2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI): . Paper presented at 2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI), Houston, TX, USA, April 11-17, 2025 (pp. 1-4).
Open this publication in new window or tab >>Structured regularization using approximate morphology for Alzheimer's disease classification
Show others...
2025 (English)In: 2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI), 2025, p. 1-4Conference paper, Published paper (Refereed)
Abstract [en]

Structured regularization allows machine learning models to consider spatial relationships among parameters, leading to results that generalize better and are more interpretable compared to norm penalties. In this study, we evaluated a novel structured regularization method that incorporates approximate morphology operators defined using harmonic mean-based fW-filters. We extended this method to multiclass classification and conducted experiments aimed at classifying magnetic resonance images (MRI) of subjects into four stages of Alzheimer's disease progression. The experimental results demonstrate that the novel structured regularization method not only performs better than standard sparse and structured regularization methods in terms of prediction accuracy (ACC), F1 scores, and the area under the receiver operating characteristic curve (AUC), but also produces interpretable coefficient maps.

Series
Proceedings (International Symposium on Biomedical Imaging), ISSN 1945-7928, E-ISSN 1945-8452
Keywords
Structured regularization, MRI, Alzheimer’s disease, Classification, Interpretation
National Category
Computer graphics and computer vision Neurosciences Artificial Intelligence
Identifiers
urn:nbn:se:umu:diva-239040 (URN)10.1109/ISBI60581.2025.10981098 (DOI)2-s2.0-105005824554 (Scopus ID)979-8-3315-2052-6 (ISBN)979-8-3315-2053-3 (ISBN)
Conference
2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI), Houston, TX, USA, April 11-17, 2025
Funder
Swedish Research Council, 2021-04810Lions Cancerforskningsfond i Norr, LP 24-2367
Available from: 2025-05-21 Created: 2025-05-21 Last updated: 2025-06-02Bibliographically approved
Lin, D., Hägg, L., Wadbro, E., Berggren, M. & Löfstedt, T. (2025). Structured regularization with object size selection using mathematical morphology. Pattern Analysis and Applications, 28, Article ID 70.
Open this publication in new window or tab >>Structured regularization with object size selection using mathematical morphology
Show others...
2025 (English)In: Pattern Analysis and Applications, ISSN 1433-7541, E-ISSN 1433-755X, Vol. 28, article id 70Article in journal (Refereed) Published
Abstract [en]

We propose a novel way to incorporate morphology operators through structured regularization of machine learning models. Specifically, we introduce a feature map in the models that performs structured variable selection. The feature map is automatically processed by approximate morphology operators and is learned together with the model coefficients. Experiments were conducted with linear regression on both synthetic data, demonstrating that the proposed methods are effective in selecting groups of parameters with much less noise than baseline models, and on three-dimensional T1-weighted brain magnetic resonance images (MRI) for age prediction, demonstrating that the proposed methods enforce sparsity and select homogeneous regions of non-zero and relevant regression coefficients. The proposed methods improve interpretability in pattern analysis. The minimum size of features in the structured variable selection can be controlled by adjusting the structuring element in the approximate morphology operator, tailored to the specific study of interest. With these added benefits, the proposed methods still perform on par with commonly used variable selection and structured variable selection methods in terms of the coefficient of determination and the Pearson correlation coefficient.

Place, publisher, year, edition, pages
Springer Nature, 2025
Keywords
Structured regularization, Approximate morphology operators, Feature selection, fW-mean filters
National Category
Artificial Intelligence Computer graphics and computer vision
Identifiers
urn:nbn:se:umu:diva-236995 (URN)10.1007/s10044-025-01444-7 (DOI)001455367400002 ()2-s2.0-105001489397 (Scopus ID)
Funder
Swedish Research Council, 2021-04810Lions Cancerforskningsfond i Norr, LP 24-2367
Available from: 2025-03-27 Created: 2025-03-27 Last updated: 2025-04-29Bibliographically approved
Meyers, C., Löfstedt, T. & Elmroth, E. (2024). Massively parallel evasion attacks and the pitfalls of adversarial retraining. EAI Endorsed Transactions on Internet of Things, 10
Open this publication in new window or tab >>Massively parallel evasion attacks and the pitfalls of adversarial retraining
2024 (English)In: EAI Endorsed Transactions on Internet of Things, E-ISSN 2414-1399, Vol. 10Article in journal (Refereed) Published
Abstract [en]

Even with widespread adoption of automated anomaly detection in safety-critical areas, both classical and advanced machine learning models are susceptible to first-order evasion attacks that fool models at run-time (e.g. an automated firewall or an anti-virus application). Kernelized support vector machines (KSVMs) are an especially useful model because they combine a complex geometry with low run-time requirements (e.g. when compared to neural networks), acting as a run-time lower bound when compared to contemporary models (e.g. deep neural networks), to provide a cost-efficient way to measure model and attack run-time costs. To properly measure and combat adversaries, we propose a massively parallel projected gradient descent (PGD) evasion attack framework. Through theoretical examinations and experiments carried out using linearly-separable Gaussian normal data, we present (i) a massively parallel naive attack, we show that adversarial retraining is unlikely to be an effective means to combat an attacker even on linearly separable datasets, (ii) a cost effective way of evaluating models defences and attacks, and an extensible code base for doing so, (iii) an inverse relationship between adversarial robustness and benign accuracy, (iv) the lack of a general relationship between attack time and efficacy, and (v) that adversarial retraining increases compute time exponentially while failing to reliably prevent highly-confident false classifications.

Place, publisher, year, edition, pages
Gent EAI, 2024
Keywords
Machine Learning, Support Vector Machines, Trustworthy AI, Anomaly Detection, AI for Cybersecurity
National Category
Computer graphics and computer vision Computer and Information Sciences
Identifiers
urn:nbn:se:umu:diva-228214 (URN)10.4108/eetiot.6652 (DOI)2-s2.0-85200255571 (Scopus ID)
Funder
Knut and Alice Wallenberg Foundation, 2019.0352
Available from: 2024-08-05 Created: 2024-08-05 Last updated: 2025-05-19Bibliographically approved
Simkó, A., Bylund, M., Jönsson, G., Löfstedt, T., Garpebring, A., Nyholm, T. & Jonsson, J. (2024). Towards MR contrast independent synthetic CT generation. Zeitschrift für Medizinische Physik, 34(2), 270-277
Open this publication in new window or tab >>Towards MR contrast independent synthetic CT generation
Show others...
2024 (English)In: Zeitschrift für Medizinische Physik, ISSN 0939-3889, E-ISSN 1876-4436, Vol. 34, no 2, p. 270-277Article in journal (Refereed) Published
Abstract [en]

The use of synthetic CT (sCT) in the radiotherapy workflow would reduce costs and scan time while removing the uncertainties around working with both MR and CT modalities. The performance of deep learning (DL) solutions for sCT generation is steadily increasing, however most proposed methods were trained and validated on private datasets of a single contrast from a single scanner. Such solutions might not perform equally well on other datasets, limiting their general usability and therefore value. Additionally, functional evaluations of sCTs such as dosimetric comparisons with CT-based dose calculations better show the impact of the methods, but the evaluations are more labor intensive than pixel-wise metrics.

To improve the generalization of an sCT model, we propose to incorporate a pre-trained DL model to pre-process the input MR images by generating artificial proton density, T1 and T2 maps (i.e. contrast-independent quantitative maps), which are then used for sCT generation. Using a dataset of only T2w MR images, the robustness towards input MR contrasts of this approach is compared to a model that was trained using the MR images directly. We evaluate the generated sCTs using pixel-wise metrics and calculating mean radiological depths, as an approximation of the mean delivered dose. On T2w images acquired with the same settings as the training dataset, there was no significant difference between the performance of the models. However, when evaluated on T1w images, and a wide range of other contrasts and scanners from both public and private datasets, our approach outperforms the baseline model. Using a dataset of T2w MR images, our proposed model implements synthetic quantitative maps to generate sCT images, improving the generalization towards other contrasts. Our code and trained models are publicly available.

Place, publisher, year, edition, pages
Elsevier, 2024
Keywords
MRI contrast, Robust machine learning, Synthetic CT generation
National Category
Computer Sciences Radiology, Nuclear Medicine and Medical Imaging
Identifiers
urn:nbn:se:umu:diva-214270 (URN)10.1016/j.zemedi.2023.07.001 (DOI)001246727700001 ()37537099 (PubMedID)2-s2.0-85169824488 (Scopus ID)
Funder
Cancerforskningsfonden i Norrland, LP 18-2182Cancerforskningsfonden i Norrland, AMP 18-912Cancerforskningsfonden i Norrland, AMP 20-1014Cancerforskningsfonden i Norrland, LP 22-2319Region VästerbottenSwedish National Infrastructure for Computing (SNIC)
Available from: 2023-09-12 Created: 2023-09-12 Last updated: 2024-07-04Bibliographically approved
Löfstedt, T. (2024). Using the Krylov subspace formulation to improve regularisation and interpretation in partial least squares regression. Computational statistics (Zeitschrift)
Open this publication in new window or tab >>Using the Krylov subspace formulation to improve regularisation and interpretation in partial least squares regression
2024 (English)In: Computational statistics (Zeitschrift), ISSN 0943-4062, E-ISSN 1613-9658Article in journal (Refereed) Epub ahead of print
Abstract [en]

Partial least squares regression (PLS-R) has been an important regression method in the life sciences and many other fields for decades. However, PLS-R is typically solved using an opaque algorithmic approach, rather than through an optimisation formulation and procedure. There is a clear optimisation formulation of the PLS-R problem based on a Krylov subspace formulation, but it is only rarely considered. The popularity of PLS-R is attributed to the ability to interpret the data through the model components, but the model components are not available when solving the PLS-R problem using the Krylov subspace formulation. We therefore highlight a simple reformulation of the PLS-R problem using the Krylov subspace formulation as a promising modelling framework for PLS-R, and illustrate one of the main benefits of this reformulation—that it allows arbitrary penalties of the regression coefficients in the PLS-R model. Further, we propose an approach to estimate the PLS-R model components for the solution found through the Krylov subspace formulation, that are those we would have obtained had we been able to use the common algorithms for estimating the PLS-R model. We illustrate the utility of the proposed method on simulated and real data.

Place, publisher, year, edition, pages
Springer Nature, 2024
Keywords
Interpretation, Krylov subspace, Partial least squares regression, PLS-R, Regularisation
National Category
Computational Mathematics Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-229897 (URN)10.1007/s00180-024-01545-7 (DOI)001310597500001 ()2-s2.0-85203694329 (Scopus ID)
Funder
Swedish Research Council, 2021-04810
Available from: 2024-09-25 Created: 2024-09-25 Last updated: 2024-09-25
Hellström, M., Löfstedt, T. & Garpebring, A. (2023). Denoising and uncertainty estimation in parameter mapping with approximate Bayesian deep image priors. Magnetic Resonance in Medicine, 90(6), 2557-2571
Open this publication in new window or tab >>Denoising and uncertainty estimation in parameter mapping with approximate Bayesian deep image priors
2023 (English)In: Magnetic Resonance in Medicine, ISSN 0740-3194, E-ISSN 1522-2594, Vol. 90, no 6, p. 2557-2571Article in journal (Refereed) Published
Abstract [en]

Purpose: To mitigate the problem of noisy parameter maps with high uncertainties by casting parameter mapping as a denoising task based on Deep Image Priors.

Methods: We extend the concept of denoising with Deep Image Prior (DIP) into parameter mapping by treating the output of an image-generating network as a parametrization of tissue parameter maps. The method implicitly denoises the parameter mapping process by filtering low-level image features with an untrained convolutional neural network (CNN). Our implementation includes uncertainty estimation from Bernoulli approximate variational inference, implemented with MC dropout, which provides model uncertainty in each voxel of the denoised parameter maps. The method is modular, so the specifics of different applications (e.g., T1 mapping) separate into application-specific signal equation blocks. We evaluate the method on variable flip angle T1 mapping, multi-echo T2 mapping, and apparent diffusion coefficient mapping.

Results: We found that deep image prior adapts successfully to several applications in parameter mapping. In all evaluations, the method produces noise-reduced parameter maps with decreased uncertainty compared to conventional methods. The downsides of the proposed method are the long computational time and the introduction of some bias from the denoising prior.

Conclusion: DIP successfully denoise the parameter mapping process and applies to several applications with limited hyperparameter tuning. Further, it is easy to implement since DIP methods do not use network training data. Although time-consuming, uncertainty information from MC dropout makes the method more robust and provides useful information when properly calibrated.

Place, publisher, year, edition, pages
John Wiley & Sons, 2023
Keywords
deep image prior, denoising, parameter mapping, quantitative MRI, uncertainty estimation
National Category
Medical Imaging Computer graphics and computer vision
Identifiers
urn:nbn:se:umu:diva-213711 (URN)10.1002/mrm.29823 (DOI)001049833500001 ()37582257 (PubMedID)2-s2.0-85168117341 (Scopus ID)
Funder
Swedish Research Council, 2019‐0432Region Västerbotten, RV‐970119Cancerforskningsfonden i Norrland, AMP 18‐912
Available from: 2023-09-14 Created: 2023-09-14 Last updated: 2025-02-09Bibliographically approved
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0001-7119-7646

Search in DiVA

Show all publications