Search

Your search keyword '"Löfstedt, P."' showing total 358 results
Start Over Author "Löfstedt, P."
358 results on '"Löfstedt, P."'

1. Compressing regularised dynamics improves link prediction in sparse networks

Author

Lindström, Maja, Blöcker, Christopher, Löfstedt, Tommy, and Rosvall, Martin

Subjects
Computer Science - Social and Information Networks
Abstract

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 incorporate a global regularisation method based on a Bayesian estimate of the transition rates along with three local regularisation methods. The regularised versions of the map equation compensate for incomplete observations and decrease the number of identified communities in sparse networks. The regularised 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 regularisation methods, the global regularisation method provides the most reliable community detection and the highest link prediction performance across different network densities.

Published
2024

2. A Cost-Aware Approach to Adversarial Robustness in Neural Networks

Author

Meyers, Charles, Sedghpour, Mohammad Reza Saleh, Löfstedt, Tommy, and Elmroth, Erik

Subjects
Computer Science - Cryptography and Security, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Statistics - Applications
Abstract

Considering the growing prominence of production-level AI and the threat of adversarial attacks that can evade a model at run-time, evaluating the robustness of models to these evasion attacks is of critical importance. Additionally, testing model changes likely means deploying the models to (e.g. a car or a medical imaging device), or a drone to see how it affects performance, making un-tested changes a public problem that reduces development speed, increases cost of development, and makes it difficult (if not impossible) to parse cause from effect. In this work, we used survival analysis as a cloud-native, time-efficient and precise method for predicting model performance in the presence of adversarial noise. For neural networks in particular, the relationships between the learning rate, batch size, training time, convergence time, and deployment cost are highly complex, so researchers generally rely on benchmark datasets to assess the ability of a model to generalize beyond the training data. To address this, we propose using accelerated failure time models to measure the effect of hardware choice, batch size, number of epochs, and test-set accuracy by using adversarial attacks to induce failures on a reference model architecture before deploying the model to the real world. We evaluate several GPU types and use the Tree Parzen Estimator to maximize model robustness and minimize model run-time simultaneously. This provides a way to evaluate the model and optimise it in a single step, while simultaneously allowing us to model the effect of model parameters on training time, prediction time, and accuracy. Using this technique, we demonstrate that newer, more-powerful hardware does decrease the training time, but with a monetary and power cost that far outpaces the marginal gains in accuracy.

Published
2024

3. A Correlation- and Mean-Aware Loss Function and Benchmarking Framework to Improve GAN-based Tabular Data Synthesis

Author

Vu, Minh H., Edler, Daniel, Wibom, Carl, Löfstedt, Tommy, Melin, Beatrice, and Rosvall, Martin

Subjects
Computer Science - Machine Learning
Abstract

Advancements in science rely on data sharing. In medicine, where personal data are often involved, synthetic tabular data generated by generative adversarial networks (GANs) offer a promising avenue. However, existing GANs struggle to capture the complexities of real-world tabular data, which often contain a mix of continuous and categorical variables with potential imbalances and dependencies. We propose a novel correlation- and mean-aware loss function designed to address these challenges as a regularizer for GANs. To ensure a rigorous evaluation, we establish a comprehensive benchmarking framework using ten real-world datasets and eight established tabular GAN baselines. The proposed loss function demonstrates statistically significant improvements over existing methods in capturing the true data distribution, significantly enhancing the quality of synthetic data generated with GANs. The benchmarking framework shows that the enhanced synthetic data quality leads to improved performance in downstream machine learning (ML) tasks, ultimately paving the way for easier data sharing., Comment: n.a

Published
2024

4. Synthesizing multi-log grasp poses

Author

Fälldin, Arvid, Wallin, Erik, Löfstedt, Tommy, and Servin, Martin

Subjects
Computer Science - Robotics
Abstract

Multi-object grasping is a challenging task. It is important for energy and cost-efficient operation of industrial crane manipulators, such as those used to collect tree logs off the forest floor and onto forest machines. In this work, we used synthetic data from physics simulations to explore how data-driven modeling can be used to infer multi-object grasp poses from images. We showed that convolutional neural networks can be trained specifically for synthesizing multi-object grasps. Using RGB-Depth images and instance segmentation masks as input, a U-Net model outputs grasp maps with corresponding grapple orientation and opening width. Given an observation of a pile of logs, the model can be used to synthesize and rate the possible grasp poses and select the most suitable one, with the possibility to respect changing operational constraints such as lift capacity and reach. When tested on previously unseen data, the proposed model found successful grasp poses with an accuracy of 95%.

Published
2024

6. A Training Rate and Survival Heuristic for Inference and Robustness Evaluation (TRASHFIRE)

Author

Meyers, Charles, Sedghpour, Mohammad Reza Saleh, Löfstedt, Tommy, and Elmroth, Erik

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Computer Vision and Pattern Recognition, Statistics - Machine Learning
Abstract

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 (\textit{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.

Published
2024

7. LatentAugment: Data Augmentation via Guided Manipulation of GAN's Latent Space

Author

Tronchin, Lorenzo, Vu, Minh H., Soda, Paolo, and Löfstedt, Tommy

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Data Augmentation (DA) is a technique to increase the quantity and diversity of the training data, and by that alleviate overfitting and improve generalisation. However, standard DA produces synthetic data for augmentation with limited diversity. Generative Adversarial Networks (GANs) may unlock additional information in a dataset by generating synthetic samples having the appearance of real images. However, these models struggle to simultaneously address three key requirements: fidelity and high-quality samples; diversity and mode coverage; and fast sampling. Indeed, GANs generate high-quality samples rapidly, but have poor mode coverage, limiting their adoption in DA applications. We propose LatentAugment, a DA strategy that overcomes the low diversity of GANs, opening up for use in DA applications. Without external supervision, LatentAugment modifies latent vectors and moves them into latent space regions to maximise the synthetic images' diversity and fidelity. It is also agnostic to the dataset and the downstream task. A wide set of experiments shows that LatentAugment improves the generalisation of a deep model translating from MRI-to-CT beating both standard DA as well GAN-based sampling. Moreover, still in comparison with GAN-based sampling, LatentAugment synthetic samples show superior mode coverage and diversity. Code is available at: https://github.com/ltronchin/LatentAugment.

Published
2023

10. Towards MR contrast independent synthetic CT generation

Author

Attila Simkó, Mikael Bylund, Gustav Jönsson, Tommy Löfstedt, Anders Garpebring, Tufve Nyholm, and Joakim Jonsson

Subjects
Synthetic CT generation, MRI contrast, Robust machine learning, Medical physics. Medical radiology. Nuclear medicine, R895-920
Abstract

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.

Published
2024
Full Text
View/download PDF

11. Reproducibility of the Methods in Medical Imaging with Deep Learning

Author

Simko, Attila, Garpebring, Anders, Jonsson, Joakim, Nyholm, Tufve, and Löfstedt, Tommy

Subjects
Computer Science - Machine Learning, Computer Science - Computer Vision and Pattern Recognition, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Concerns about the reproducibility of deep learning research are more prominent than ever, with no clear solution in sight. The relevance of machine learning research can only be improved if we also employ empirical rigor that incorporates reproducibility guidelines, especially so in the medical imaging field. The Medical Imaging with Deep Learning (MIDL) conference has made advancements in this direction by advocating open access, and recently also recommending authors to make their code public - both aspects being adopted by the majority of the conference submissions. This helps the reproducibility of the methods, however, there is currently little or no support for further evaluation of these supplementary material, making them vulnerable to poor quality, which affects the impact of the entire submission. We have evaluated all accepted full paper submissions to MIDL between 2018 and 2022 using established, but slightly adjusted guidelines on reproducibility and the quality of the public repositories. The evaluations show that publishing repositories and using public datasets are becoming more popular, which helps traceability, but the quality of the repositories has not improved over the years, leaving room for improvement in every aspect of designing repositories. Merely 22% of all submissions contain a repository that were deemed repeatable using our evaluations. From the commonly encountered issues during the evaluations, we propose a set of guidelines for machine learning-related research for medical imaging applications, adjusted specifically for future submissions to MIDL.

Published
2022

12. Associations between dietary behaviours and the mental and physical well-being of Swedish adolescents

Author

Kenisha Russell Jonsson, Cameron Kymani Bailey, Maria Corell, Petra Löfstedt, and Nicholas Kofi Adjei

Subjects
Eating behaviour, Diet, Mental health, Physical health, Socioeconomic disparities, Adolescents, Pediatrics, RJ1-570, Psychiatry, RC435-571
Abstract

Abstract Aims This study aims to investigate the association between dietary behaviours, overweight/obesity, and mental health and well-being among Swedish adolescents. Methods Data from the 2017/2018 Health Behaviour in School-aged Children (HBSC) survey of 3692 adolescents aged ≈11, ≈13, and ≈15 years was analysed. We evaluated the prevalence and association of dietary behaviours, characterised by daily intake of fruits and vegetables, consumption of sugar-sweetened beverages (SSBs) and sweets once per week, and daily consumption of family meals and breakfast, with socioeconomic and demographic factors. Multivariate logistic regression models (adjusted and unadjusted) were then used to examine the relationship between dietary behaviours, overweight/obesity, and mental health and well-being including psychosomatic complaints, life satisfaction, and school-related pressure. Results Boys were more likely to eat breakfast and have meals together with their family than girls, but their fruit and vegetable intake was lower compared to girls. Adolescents with lower socioeconomic status (low family affluence, families perceived ’not well-off’ and two unemployed parents), single-parent households and twoforeign born parents were more likely to engage in poorer dietary behaviours. Daily breakfast, family meals, fruit and vegetable intake were positively associated with mental health and well-being. Specifically, daily family meals were linked to higher life satisfaction, fewer psychosomatic complaints, and reduced school-related pressure. Breakfast emerged as a beneficial dietary habit, associated with higher life satisfaction, and a lower likelihood of psychosomatic complaints, school-related pressure, and overweight/obesity. Fruit and vegetable consumption was consistently associated with better mental health and well-being while associations between SSBs and sweets were mixed. Conclusions Improving mental health and well-being, along with tackling the rising rates of mental illness and challenges related to overweight/obesity in adolescents constitute key public health priorities. Implementing policies that promote the intake of fruits and vegetables, reducing the consumption of SSBs and sweets, and emphasizing the value of having breakfast and sharing family meals could offer a cost-effective public health intervention.

Published
2024
Full Text
View/download PDF

13. Socioeconomic inequalities in adolescent mental health in the Nordic countries in the 2000s - A study using cross-sectional data from the Health Behaviour in School-aged Children study

Author

Maria Corell, Peter Friberg, Max Petzold, and Petra Löfstedt

Subjects
Adolescent health, Mental health, Subjective health complaints, Life satisfaction, Income inequality, At-risk-of-poverty rate, Public aspects of medicine, RA1-1270
Abstract

Abstract Background Adolescents in Sweden experience more mental health problems and lower mental well-being than adolescents in other Nordic countries. According to the literature, one possible explanation may be differences in income inequality. The at-risk-of-poverty rate varies significantly across the Nordic countries, and the highest rate is found in Sweden. The aims of the study were to examine socioeconomic inequalities in subjective health complaints and life satisfaction among adolescents in the Nordic countries during 2002 − 2018 and to explore whether subjective health complaints and life satisfaction were related to income inequality in terms of the at-risk-of-poverty rate at the country level. Methods Data regarding 15-year-olds from the Health Behaviour in School-aged Children study from five survey rounds (2002 − 2018) were used (n = 41,148). The HBSC Symptoms Checklist and Cantril’s ladder were used as measures of subjective health complaints and life satisfaction, respectively. The Family Affluence Scale, the Perceived Family Wealth item and the at-risk-of-poverty rate in each country were used as measures of individual-level socioeconomic conditions and country-level income inequality. Statistical methods involved ANOVA, multiple linear regressions and multilevel regression analyses. Results Absolute and relative socioeconomic inequalities in both subjective health complaints and life satisfaction were found in all countries. Sweden showed average socioeconomic inequalities, Iceland the largest and Denmark the smallest. Country-level income inequality in terms of the at-risk-of-poverty rate was associated with a higher prevalence of subjective health complaints and lower levels of life satisfaction in all countries. Conclusion Socioeconomic inequalities in adolescent mental health and well-being persisted in Nordic countries in the 2000s. Increasing income inequality may have contributed to higher levels of SHC and lower LS in Sweden compared to the other Nordic countries. Policies improving families’ socioeconomic conditions and reducing income inequality at the country level are needed to improve and reduce inequalities in mental health and well-being among adolescents.

Published
2024
Full Text
View/download PDF

16. Improving MR image quality with a multi-task model, using convolutional losses

Author

Attila Simkó, Simone Ruiter, Tommy Löfstedt, Anders Garpebring, Tufve Nyholm, Mikael Bylund, and Joakim Jonsson

Subjects
Machine learning, Magnetic resonance imaging, Image artefact correction, Medical technology, R855-855.5
Abstract

Abstract Purpose During the acquisition of MRI data, patient-, sequence-, or hardware-related factors can introduce artefacts that degrade image quality. Four of the most significant tasks for improving MRI image quality have been bias field correction, super-resolution, motion-, and noise correction. Machine learning has achieved outstanding results in improving MR image quality for these tasks individually, yet multi-task methods are rarely explored. Methods In this study, we developed a model to simultaneously correct for all four aforementioned artefacts using multi-task learning. Two different datasets were collected, one consisting of brain scans while the other pelvic scans, which were used to train separate models, implementing their corresponding artefact augmentations. Additionally, we explored a novel loss function that does not only aim to reconstruct the individual pixel values, but also the image gradients, to produce sharper, more realistic results. The difference between the evaluated methods was tested for significance using a Friedman test of equivalence followed by a Nemenyi post-hoc test. Results Our proposed model generally outperformed other commonly-used correction methods for individual artefacts, consistently achieving equal or superior results in at least one of the evaluation metrics. For images with multiple simultaneous artefacts, we show that the performance of using a combination of models, trained to correct individual artefacts depends heavily on the order that they were applied. This is not an issue for our proposed multi-task model. The model trained using our novel convolutional loss function always outperformed the model trained with a mean squared error loss, when evaluated using Visual Information Fidelity, a quality metric connected to perceptual quality. Conclusion We trained two models for multi-task MRI artefact correction of brain, and pelvic scans. We used a novel loss function that significantly improves the image quality of the outputs over using mean squared error. The approach performs well on real world data, and it provides insight into which artefacts it detects and corrects for. Our proposed model and source code were made publicly available.

Published
2023
Full Text
View/download PDF

17. A Data-Adaptive Loss Function for Incomplete Data and Incremental Learning in Semantic Image Segmentation

Author

Vu, Minh H., Norman, Gabriella, Nyholm, Tufve, and Löfstedt, Tommy

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition
Abstract

In the last years, deep learning has dramatically improved the performances in a variety of medical image analysis applications. Among different types of deep learning models, convolutional neural networks have been among the most successful and they have been used in many applications in medical imaging. Training deep convolutional neural networks often requires large amounts of image data to generalize well to new unseen images. It is often time-consuming and expensive to collect large amounts of data in the medical image domain due to expensive imaging systems, and the need for experts to manually make ground truth annotations. A potential problem arises if new structures are added when a decision support system is already deployed and in use. Since the field of radiation therapy is constantly developing, the new structures would also have to be covered by the decision support system. In the present work, we propose a novel loss function, that adapts to the available data in order to utilize all available data, even when some have missing annotations. We demonstrate that the proposed loss function also works well in an incremental learning setting, where it can automatically incorporate new structures as they appear. Experiments on a large in-house data set show that the proposed method performs on par with baseline models, while greatly reducing the training time.

Published
2021

18. Multi-Decoder Networks with Multi-Denoising Inputs for Tumor Segmentation

Author

Vu, Minh H., Nyholm, Tufve, and Löfstedt, Tommy

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition
Abstract

Automatic segmentation of brain glioma from multimodal MRI scans plays a key role in clinical trials and practice. Unfortunately, manual segmentation is very challenging, time-consuming, costly, and often inaccurate despite human expertise due to the high variance and high uncertainty in the human annotations. In the present work, we develop an end-to-end deep-learning-based segmentation method using a multi-decoder architecture by jointly learning three separate sub-problems using a partly shared encoder. We also propose to apply smoothing methods to the input images to generate denoised versions as additional inputs to the network. The validation performance indicate an improvement when using the proposed method. The proposed method was ranked 2nd in the task of Quantification of Uncertainty in Segmentation in the Brain Tumors in Multimodal Magnetic Resonance Imaging Challenge 2020.

Published
2020

19. A Question-Centric Model for Visual Question Answering in Medical Imaging

Author

Vu, Minh H., Löfstedt, Tommy, Nyholm, Tufve, and Sznitman, Raphael

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Image and Video Processing, Statistics - Machine Learning
Abstract

Deep learning methods have proven extremely effective at performing a variety of medical image analysis tasks. With their potential use in clinical routine, their lack of transparency has however been one of their few weak points, raising concerns regarding their behavior and failure modes. While most research to infer model behavior has focused on indirect strategies that estimate prediction uncertainties and visualize model support in the input image space, the ability to explicitly query a prediction model regarding its image content offers a more direct way to determine the behavior of trained models. To this end, we present a novel Visual Question Answering approach that allows an image to be queried by means of a written question. Experiments on a variety of medical and natural image datasets show that by fusing image and question features in a novel way, the proposed approach achieves an equal or higher accuracy compared to current methods., Comment: Accepted at IEEE Transactions on Medical Imaging

Published
2020
Full Text
View/download PDF

20. Why Did Jesus’ Disciples Fail to Cast Out the Deaf and Mute Spirit? (Mark 9:14–29)

Author

Torsten Löfstedt

Subjects
exorcism, demons, spirits, transfiguration, Mark 9:29, raphael, The Bible, BS1-2970
Abstract

On the basis of a close reading of the pericope of the healing of the boy with a mute spirit as it is recorded in Mark 9:14–29, I argue that Mark intended Jesus’ enigmatic answer “this kind can can come out only through prayer and fasting” (Mark 9:29, NRSV) to be a lesson for his readers on how to prepare for exorcisms. Like the disciples who failed to cast out the mute demon while Jesus’ was being transfigured, so too Mark’s readers must prepare for exorcisms by prayer and fasting, for Jesus is not present with them the way he was for his disciples during most of his earthly ministry.

Published
2023
Full Text
View/download PDF

21. Collaborative challenges and barriers when planning and implementing Bus Rapid Transit (BRT). Lessons from Swedish BRT projects

Author

Jakob Allansson, Fredrik Pettersson-Löfstedt, and Robert Hrelja

Subjects
Bus rapid transit, BRT, collaboration, barriers, planning, implementation, City planning, HT165.5-169.9, Transportation and communications, HE1-9990
Abstract

ABSTRACTThe aim of this paper is to improve the knowledge of collaborative challenges when planning and implementing Bus Rapid Transit (BRT). Collaborative challenges are here understood as the barriers that may arise in BRT planning and implementation as a consequence of several formally independent actors, occasionally with different interests, participating in the planning. The results are based on an analysis of actor interactions in Swedish BRT projects. These projects are analysed in relation to the state of the art in the research field of collaborative approaches. The results show two main and interrelated collaborative challenges. The first category of challenges concerns difficulties for actors in creating a common understanding of what a BRT system is, the second category concerns details of bus priority measures, e.g. busways, priority at intersections, and how to handle and deal with conflicting interests when removing speed bumps or pedestrian and cycle crossings. In terms of policy is in the early stages of the planning processes. This can be generated by working practices and tools that facilitate agreements on how to handle different interests and trade-offs. BRT guidelines adapted to national transport policy, legal and organisational conditions could function as tools in assisting actor dialogue.

Published
2023
Full Text
View/download PDF

22. Evaluation of Multi-Slice Inputs to Convolutional Neural Networks for Medical Image Segmentation

Author

Vu, Minh H., Grimbergen, Guus, Nyholm, Tufve, and Löfstedt, Tommy

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Statistics - Machine Learning
Abstract

When using Convolutional Neural Networks (CNNs) for segmentation of organs and lesions in medical images, the conventional approach is to work with inputs and outputs either as single slice (2D) or whole volumes (3D). One common alternative, in this study denoted as pseudo-3D, is to use a stack of adjacent slices as input and produce a prediction for at least the central slice. This approach gives the network the possibility to capture 3D spatial information, with only a minor additional computational cost. In this study, we systematically evaluate the segmentation performance and computational costs of this pseudo-3D approach as a function of the number of input slices, and compare the results to conventional end-to-end 2D and 3D CNNs. The standard pseudo-3D method regards the neighboring slices as multiple input image channels. We additionally evaluate a simple approach where the input stack is a volumetric input that is repeatably convolved in 3D to obtain a 2D feature map. This 2D map is in turn fed into a standard 2D network. We conducted experiments using two different CNN backbone architectures and on five diverse data sets covering different anatomical regions, imaging modalities, and segmentation tasks. We found that while both pseudo-3D methods can process a large number of slices at once and still be computationally much more efficient than fully 3D CNNs, a significant improvement over a regular 2D CNN was only observed for one of the five data sets. An analysis of the structural properties of the segmentation masks revealed no relations to the segmentation performance with respect to the number of input slices. The conclusion is therefore that in the general case, multi-slice inputs appear to not significantly improve segmentation results over using 2D or 3D CNNs.

Published
2019
Full Text
View/download PDF

26. Book Reviews

Author

Carl Johan Berglund, LarsOlov Eriksson, Stefan Green, Gunnar Haaland, Mikael Larsson, Hans Leander, Torsten Löfstedt, Dan Nässelqvist, Adam Sabir, Karl Olav Sandnes, Hanna Stenström, Mikael Tellbe, Lena-Sofia Tiemeyer, Cecilia Wassén, Tommy Wasserman, Martin Wessbrandt, Ola Wikander, David Willgren, and Tobias Ålöw

Subjects
The Bible, BS1-2970
Abstract

The following books are reviewed: Heinrich Assell, Stefan Beyerle och Christfried Böttrich (red.), Beyond Biblical Theologies, (Hanna Stenström) Jean-Dominique Barthélemy, Studies in the Text of the Old Testament: An Introduction to the Hebrew Old Testament Text Project (LarsOlov Eriksson) Brennan W. Breed, Nomadic Texts: A Theory of Biblical Reception History (Mikael Larsson) Walter Brueggemann och William H. Bellinger Jr, Psalms (David Willgren) Sean Burt, The Courtier and the Governor: Transformation of Genre in the Nehemiah Memoir (Lena-Sofia Tiemeyer) Chrys C. Caragounis, New Testament Language and Exegesis: A Diachronic Approach (Dan Nässelqvist) Cavan W. Concannon, “When You Were Gentiles”: Specters of Ethnicity in Roman Corinth and Paul's Corinthian Correspondence (Adam Sabir) Ovidiu Creangă (red.), Men and Masculinity in the Hebrew Bible and Beyond (Mikael Larsson) Ovidiu Creangă och Peter-Ben Smit (red.), Biblical Masculinites Foregrounded (Mikael Larsson) Roland Deines, Acts of God in History: Studies Towards Recovering a Theological Historiography (Tobias Ålöw) Anders Ekenberg, Jonas Holmstrand och Mikael Winninge (red.), 2000 år med Paulus (Hans Leander) Magnus Evertsson, Liknelser och läsningar: Reception av liknelseberättelser ur Lukasevangeliet, kapitel 10–15, i predikoutkast för Svenska kyrkan 1985–2013 (Karl Olav Sandnes) Josef Forsling, Composite Artistry in Numbers: A Study in Biblical Narrative Conventions (Ola Wikander) Ida Fröhlich and Erkki Koskenniemi (eds.), Evil and the Devil (Torsten Löfstedt) Susan Gillingham, A Journey of Two Psalms: The Reception of Psalms 1 and 2 in Jewish and Christian Tradition (David Willgren) Raimo Hakola, Nina Nikki and Ulla Tervahauta (eds.), Others and the Construction of Early Christian Identities (Cecilia Wassén) Thomas Hieke and Tobias Nicklas (eds.), The Day of Atonement: Its Interpretation in Early Jewish and Christian Traditions (Lena-Sofia Tiemeyer) Yulin Liu, Temple Purity in 1–2 Corinthians (Martin Wessbrandt) Peter W. Martens, Origen and Scripture: The Contours of the Exegetical Life (Carl Johan Berglund) Gerson Lameck Mgaya, Spiritual Gifts: A Sociorhetorical Interpretation of 1 Cor 12–14 (Mikael Tellbe) Takayoshi Oshima, Babylonian Poems of Pious Sufferers (Lena-Sofia Tiemeyer) Stanley E. Porter och Eckhard J. Schnabel (red.), On the Writing of New Testament Commentaries: Festschrift for Grant R. Osborne on the Occasion of His 70th Birthday (LarsOlov Eriksson) Jacques T. A. G. M. van Ruiten, Abraham in the Book of Jubilees: The Rewriting of Genesis 11:26–25:10 in the Book of Jubilees 11:14–23:8 (Stefan Green) C. A. Strine, Sworn Enemies: The Divine Oath, the Book of Ezekiel, and the Polemic of Exile (Lena-Sofia Tiemeyer) Michel Tuval, From Jerusalem Priest to Roman Jew: On Josephus and the Paradigms of Ancient Judaism (Gunnar Haaland) Markus Vinzent, Marcion and the Dating of the Synoptic Gospels (Martin Wessbrandt) Georg A. Walser, Old Testament Quotations in Hebrews: Studies in their Textual and Contextual Background (Tommy Wasserman) Ellen White, Yahweh’ s Council: Its Structure and Membership (Lena-Sofia Tiemeyer) Al Wolters, Zechariah (Lena-Sofia Tiemeyer) Lorne R. Zelyck, John among the Other Gospels: The Reception of the Fourth Gospel in the Extra-Canonical Gospels (Carl Johan Berglund)

Published
2023
Full Text
View/download PDF

28. Book Reviews

Author

Anja Klein, Bo Krister Ljungberg, Carl Johan Berglund, Cecilia Wassén, Cian Power, Jan Retsö, Joel Kuhlin, Josef Forsling, LarsOlov Eriksson, Magnus Halle, Martin Wessbrandt, Mikael Larsson, Mikael Tellbe, Paul Linjamaa, Stefan Green, Stig Norin, Tobias Ålöw, and Torsten Löfstedt

Subjects
The Bible, BS1-2970
Abstract

The following books are reviewed: Aichele, George, Simulating Jesus: Reality Effects in the Gospels (Joel Kuhlin) Amos, Roger, Hypocrites or Heroes? The Paradoxical Portrayal of the Pharisees in the New Testament (Tobias Ålöw) Collins, John J., Encounters with Biblical Theology (Stig Norin) Dochhorn Jan, Susanne Rudnig-Zelt, and Benjamin Wold (eds.), Das Böse, der Teufel und Dämonen – Evil, the Devil, and Demons (Torsten Löfstedt) Ehrman, Bart D., Jesus Before the Gospels: How the Εarliest Christians Remembered, Changed, and Invented Their Stories of the Savior (Joel Kuhlin) England Emma och William John Lyons (red.), Reception History and Biblical Studies: Theory and Practice (Mikael Larsson) Fewell, Danna Nolan (ed.), The Oxford Handbook of Biblical Narrative (Josef Forsling) Gordon, Robert P. and Hans M. Barstad (eds.), “Thus speaks Ishtar of Arbela”: Prophecy in Israel, Assyria and Egypt in the Neo-Assyrian Period (Magnus Halle) Giuntoli Federico and Konrad Schmid (eds.), The Post-Priestly Pentateuch: New Perspectives on Its Redactional Development and Theological Profiles (Jan Retsö) Hayes, Elizabeth R. och Lena-Sofia Tiemeyer (red.), “I Lifted My Eyes and Saw”: Reading Dream and Vision Reports in the Hebrew Bible (Stefan Green) Heilig, Christoph, Hidden Criticism? The Methodology and Plausibilituy of the Search for a Counter-Imperial Subtext in Paul (Joel Kuhlin) Kim, Yeong Seon, The Temple Administration and the Levites in Chronicles (Jan Retsö) Klein, Anja, Geschichte und Gebet: Die Rezeption der biblischen Geschichte in den Psalmen des Alten Testaments (LarsOlov Eriksson) Klink III, Edward W., och Darian R. Lockett. Understanding Biblical Theology: A Comparison of Theory and Practice (Bo Krister Ljungberg) Knoppers, Gary N., Jews and Samaritans: The Origins and History of Their Early Relations (Cecilia Wassén) Markschies, Christoph, Christian Theology and Its Institutions in the Early Roman Empire: Prolegomena to a History of Early Christian Theology (Carl Johan Berglund) Mettinger, Tryggve N. D., Reports from a Scholar’s Life: Select Papers on the Hebrew Bible (Cian Power) Neusner, Jacob och Alan J. Avery Peck (red.); William Scott Green och Günter Stemberger (rådgivande red.), Encyclopedia of Midrash. Biblical Interpretation in Formative Judaism. Volume I-II (Tobias Ålöw) Porter, Stanley E. och David I. Yoon (red.), Paul and Gnosis (Paul Linjamaa) Smith, Geoffrey S., Guilt By Association: Heresy Catalogues in Early Christianity (Martin Wessbrandt) Strauss, Mark L. och Paul E. Engle (red.), Remarriage After Divorce in Today’s Church (Bo Krister Ljungberg) Willgren, David, The Formation of the ‘Book’ of Psalms: Reconsidering the Transmission and Canonization of Psalmody in Light of Material Culture and the Poetics of Anthologies (Anja Klein) Wright, Tom, The Day the Revolution Began: Rethinking the Meaning of Jesus’ Crucifixion (Mikael Tellbe)

Published
2023
Full Text
View/download PDF

30. End-to-End Cascaded U-Nets with a Localization Network for Kidney Tumor Segmentation

Author

Vu, Minh H., Grimbergen, Guus, Simkó, Attila, Nyholm, Tufve, and Löfstedt, Tommy

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Machine Learning, Statistics - Machine Learning
Abstract

Kidney tumor segmentation emerges as a new frontier of computer vision in medical imaging. This is partly due to its challenging manual annotation and great medical impact. Within the scope of the Kidney Tumor Segmentation Challenge 2019, that is aiming at combined kidney and tumor segmentation, this work proposes a novel combination of 3D U-Nets---collectively denoted TuNet---utilizing the resulting kidney masks for the consecutive tumor segmentation. The proposed method achieves a S{\o}rensen-Dice coefficient score of 0.902 for the kidney, and 0.408 for the tumor segmentation, computed from a five-fold cross-validation on the 210 patients available in the data., Comment: 2019 Kidney Tumor Segmentation Challenge

Published
2019
Full Text
View/download PDF

31. TuNet: End-to-end Hierarchical Brain Tumor Segmentation using Cascaded Networks

Author

Vu, Minh H., Nyholm, Tufve, and Löfstedt, Tommy

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Statistics - Machine Learning
Abstract

Glioma is one of the most common types of brain tumors; it arises in the glial cells in the human brain and in the spinal cord. In addition to having a high mortality rate, glioma treatment is also very expensive. Hence, automatic and accurate segmentation and measurement from the early stages are critical in order to prolong the survival rates of the patients and to reduce the costs of the treatment. In the present work, we propose a novel end-to-end cascaded network for semantic segmentation that utilizes the hierarchical structure of the tumor sub-regions with ResNet-like blocks and Squeeze-and-Excitation modules after each convolution and concatenation block. By utilizing cross-validation, an average ensemble technique, and a simple post-processing technique, we obtained dice scores of 88.06, 80.84, and 80.29, and Hausdorff Distances (95th percentile) of 6.10, 5.17, and 2.21 for the whole tumor, tumor core, and enhancing tumor, respectively, on the online test set., Comment: Accepted at MICCAI BrainLes 2019

Published
2019
Full Text
View/download PDF

32. A Generalized Network for MRI Intensity Normalization

Author

Simkó, A., Löfstedt, T., Garpebring, A., Nyholm, T., and Jonsson, J.

Subjects
Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Image normalization, the correction for intra-volume inhomogeneities in magnetic resonance imaging (MRI) data has little significance for visual diagnosis, but is a crucial step before automated radiotherapy solutions. There are several well-established normalization methods, however they are usually time expensive and difficult to tune for a specific dataset. In this study, we show how an artificial neural network (ANN) can be trained on non-medical images --- making the model general --- for intensity normalization on medical MRI images. Compared to one of the most well-known correction methods, N4ITK, the trained network achieves a higher accuracy with a speedup-factor of almost 70., Comment: MIDL 2019 [arXiv:1907.08612]

Published
2019

34. The clustering of multiple health and lifestyle behaviors among Swedish adolescents: a person-oriented analysis

Author

Kenisha Russell Jonsson, Maria Corell, Petra Löfstedt, and Nicholas Kofi Adjei

Subjects
health behaviors, lifestyle, alcohol, diet, physical activity, smoking, Public aspects of medicine, RA1-1270
Abstract

BackgroundKnowledge of the distribution, prevalence, and clustering of multiple health and lifestyle related behaviors (HLBs) among adolescents can inform the development of effective health-promoting policies and interventions. We assessed the clustering of multiple HLBs among 11, 13 and 15-year-old Swedish adolescents and examined the socioeconomic and demographic correlates for the identified clusters.MethodsWe used data from the 2017/2018 Swedish Health Behaviour in School-aged children (HBSC) study to conduct sex and age-stratified latent class analysis (LCA). The LCA was based on five HLBs: eating behavior and habits (EBH), physical activity (PA), tobacco usage (TU), alcohol consumption (AC) and sleeping habits and patterns (SHPs). Multinomial logistic regression models were used to assess the associations between the identified clusters and the socioeconomic and demographic characteristics of adolescents and their parents.ResultsHealth behaviors varied by sex and age. Four distinct clusters were identified based on sex: cluster 1 (Mixed eating behaviors and habits, physical activity and low alcohol consumption), cluster 2 (Healthy lifestyle behaviors), cluster 3 (Unhealthy lifestyle behaviors), and cluster 4 (Breakfast, low alcohol consumption and tobacco usage). In the age-stratified analyzes, three clusters were identified: cluster 1 (Unhealthy lifestyle behaviors), cluster 2 (Moderately healthy lifestyle behaviors) and cluster 3 (Healthy lifestyle behaviors). The multinomial analysis showed that sex, age, family situation and perceived family wealth were strong predictors of health behaviors. Unhealthy behaviors were most commonly associated with socioeconomic disadvantage, having a migrant background, and living in reconstructed families or single-parent households.ConclusionHealth behaviors vary significantly based on socioeconomic and demographic circumstances. Targeted policies and intervention programs are necessary to improve HLBs among vulnerable and at-risk adolescents.

Published
2023
Full Text
View/download PDF

35. Challenges of delivering TOD in low-density contexts: the Swedish experience of barriers and enablers

Author

Robert Hrelja, Lina Olsson, Fredrik Pettersson-Löfstedt, and Tom Rye

Subjects
Transit-oriented development, Public transport, Sustainable urban development, Low-density, Sweden, Transportation engineering, TA1001-1280, Transportation and communications, HE1-9990
Abstract

Abstract Background Transit Oriented Development (TOD) is key to the success of public transport and for sustainable urban and regional development. Previous research has often focused on delivering TOD in urban areas with high population and building densities. This highlights the need to broaden the scope of TOD approaches to also include less densely populated areas located outside the immediate urban cores as a key concern for policy. Purpose The aim of this paper is to increase the knowledge of how to deliver TOD in such low-density contexts. Methods Three case studies of attempts at delivering TOD in sparsely populated areas in three Swedish city-regions are made. The data for the case studies consist of planning documents and 13 interviews with key stakeholders involved in the planning processes of the three cases. Results The results show that many of the barriers and enablers are rather similar to those identified in research on TOD in much more urban contexts in other parts of the world, but the relationships between them are differently nuanced in low-density contexts. The lack of clear quantified definitions of what TOD is (or is not) allows a more flexible, site-specific understanding of TOD to emerge in this context. Results It is important that a shared vision of TOD in each location is developed by the organizations involved—and such a shared vision appears to be crucial for the development to have a good chance of being delivered. Informal definitions, and individual perceptions (including those of the public) are important; and in locations that are quite “marginal” for TOD, all enablers must interact together positively for the development to have the maximum likelihood of going ahead as planned.

Published
2022
Full Text
View/download PDF

36. Patronage effects of off-peak service improvements in regional public transport

Author

Joel Hansson, Fredrik Pettersson-Löfstedt, Helena Svensson, and Anders Wretstrand

Subjects
Public transport, Rural accessibility, Regional bus, Regional rail, Timetable planning, Service hours, Transportation engineering, TA1001-1280, Transportation and communications, HE1-9990
Abstract

Abstract The purpose of this study is to look into patronage effects of extended supply outside peak hours on regional public transport services. Previous studies have shown that the service frequency is an attribute of great importance for regional passengers, but little is known about the details regarding peak and off-peak frequencies or service hours. The present study addresses this knowledge gap, departing from the hypothesis that additional off-peak supply means more flexibility for the passengers in terms of departure time options, and possibly also a sense of security for passengers who are uncertain about the time of their (return) trips. Essentially, the added off-peak departures may attract more passengers even if they normally do not or only occasionally use the off-peak services. The patronage effects are explored through four case studies from the region of Scania in southern Sweden. The cases include regional rail and bus services where substantial improvements have been made outside peak hours, resulting in at least hourly all-day services. The results of the study provide new insights into the fundamental planning policy trade-off between maximum frequency and span of public transport services in urban peripheries and rural areas. Importantly, the results suggest that improved time coverage may lead to substantial patronage growth, and this growth is evident also during peak hours, despite unaltered peak hour frequencies. Hence, off-peak departures cannot be assessed only through patronage levels in isolated time periods, let alone on the individual departures.

Published
2022
Full Text
View/download PDF

37. Characterization of olfactory sensory neurons in the striped ambrosia beetle Trypodendron lineatum

Author

Twinkle Biswas, Jothi Kumar Yuvaraj, Bill S. Hansson, Christer Löfstedt, Olle Anderbrant, and Martin N. Andersson

Subjects
aggregation pheromone, fungal symbiont, olfactory sensory neuron (OSN), single sensillum recordings (SSR), mutualism, Physiology, QP1-981
Abstract

Introduction: The striped ambrosia beetle Trypodendron lineatum (Coleoptera, Curculionidae, Scolytinae) is a major forest pest in the Holarctic region. It uses an aggregation pheromone and host and non-host volatiles to locate suitable host trees, primarily stressed or dying conifer trees. The beetles bore into the xylem and inoculate spores of their obligate fungal mutualist Phialophoropsis ferruginea inside their excavated egg galleries, with the fungus serving as the main food source for the developing larvae. Olfactory sensory neuron (OSN) responses to pheromones and host volatiles are poorly understood in T. lineatum and other ambrosia beetles, and nothing is known about potential responses to fungal volatiles.Methods: We screened responses of OSNs present in 170 antennal olfactory sensilla using single sensillum recordings (SSR) and 57 odor stimuli, including pheromones, host and non-host compounds, as well as volatiles produced by P. ferruginea and fungal symbionts of other scolytine beetles.Results and Discussion: Thirteen OSN classes were characterized based on their characteristic response profiles. An OSN class responding to the aggregation pheromone lineatin was clearly the most abundant on the antennae. In addition, four OSN classes responded specifically to volatile compounds originating from the obligate fungal mutualist and three responded to non-host plant volatiles. Our data also show that T. lineatum has OSN classes tuned to pheromones of other bark beetles. Several OSN classes showed similar response profiles to those previously described in the sympatric bark beetle Ips typographus, which may reflect their shared ancestry.

Published
2023
Full Text
View/download PDF

42. Release of moth pheromone compounds from Nicotiana benthamiana upon transient expression of heterologous biosynthetic genes

Author

Yi-Han Xia, Bao-Jian Ding, Shuang-Lin Dong, Hong-Lei Wang, Per Hofvander, and Christer Löfstedt

Subjects
Functional characterization, Fatty acyl desaturases, Fatty acyl reductase, Alcohol oxidation, Acetyltransferase, Heterologous expression systems, Biology (General), QH301-705.5
Abstract

Abstract Background Using genetically modified plants as natural dispensers of insect pheromones may eventually become part of a novel strategy for integrated pest management. Results In the present study, we first characterized essential functional genes for sex pheromone biosynthesis in the rice stem borer Chilo suppressalis (Walker) by heterologous expression in Saccharomyces cerevisiae and Nicotiana benthamiana, including two desaturase genes CsupYPAQ and CsupKPSE and a reductase gene CsupFAR2. Subsequently, we co-expressed CsupYPAQ and CsupFAR2 together with the previously characterized moth desaturase Atr∆11 in N. benthamiana. This resulted in the production of (Z)-11-hexadecenol together with (Z)-11-hexadecenal, the major pheromone component of C. suppressalis. Both compounds were collected from the transformed N. benthamiana headspace volatiles using solid-phase microextraction. We finally added the expression of a yeast acetyltransferase gene ATF1 and could then confirm also (Z)-11-hexadecenyl acetate release from the plant. Conclusions Our results pave the way for stable transformation of plants to be used as biological pheromone sources in different pest control strategies.

Published
2022
Full Text
View/download PDF

43. Ionotropic receptors in the turnip moth Agrotis segetum respond to repellent medium-chain fatty acids

Author

Xiao-Qing Hou, Dan-Dan Zhang, Daniel Powell, Hong-Lei Wang, Martin N. Andersson, and Christer Löfstedt

Subjects
Xenopus oocytes, Fluorescence in situ hybridization, Acid-sensing, Repellent, Octanoic acid, Lepidoptera, Biology (General), QH301-705.5
Abstract

Abstract Background In insects, airborne chemical signals are mainly detected by two receptor families, odorant receptors (ORs) and ionotropic receptors (IRs). Functions of ORs have been intensively investigated in Diptera and Lepidoptera, while the functions and evolution of the more ancient IR family remain largely unexplored beyond Diptera. Results Here, we identified a repertoire of 26 IRs from transcriptomes of female and male antennae, and ovipositors in the moth Agrotis segetum. We observed that a large clade formed by IR75p and IR75q expansions is closely related to the acid-sensing IRs identified in Diptera. We functionally assayed each of the five AsegIRs from this clade using Xenopus oocytes and found that two receptors responded to the tested ligands. AsegIR75p.1 responded to several compounds but hexanoic acid was revealed to be the primary ligand, and AsegIR75q.1 responded primarily to octanoic acid, and less so to nonanoic acid. It has been reported that the C6-C10 medium-chain fatty acids repel various insects including many drosophilids and mosquitos. We show that the C6-C10 medium-chain fatty acids elicited antennal responses of both sexes of A. segetum, while only octanoic acid had repellent effect to the moths in a behavioral assay. In addition, using fluorescence in situ hybridization, we demonstrated that the five IRs and their co-receptor AsegIR8a are not located in coeloconic sensilla as found in Drosophila, but in basiconic or trichoid sensilla. Conclusions Our results significantly expand the current knowledge of the insect IR family. Based on the functional data in combination with phylogenetic analysis, we propose that subfunctionalization after gene duplication plays an important role in the evolution of ligand specificities of the acid-sensing IRs in Lepidoptera.

Published
2022
Full Text
View/download PDF

44. Structure of an antennally-expressed carboxylesterase suggests lepidopteran odorant degrading enzymes are broadly tuned

Author

Jacob A. Corcoran, Cyril Hamiaux, Nicoletta Faraone, Christer Löfstedt, and Colm Carraher

Subjects
Epiphyas postvittana, Carboxyl/cholinesterase, Odorant degrading enzyme, RNAseq, GC-MS, Crystal structure, Zoology, QL1-991
Abstract

Insects rely on the detection of chemical cues present in the environment to guide their foraging and reproductive behaviour. As such, insects have evolved a sophisticated chemical processing system in their antennae comprised of several types of olfactory proteins. Of these proteins, odorant degrading enzymes are responsible for metabolising the chemical cues within the antennae, thereby maintaining olfactory system function. Members of the carboxyl/cholinesterase gene family are known to degrade odorant molecules with acetate-ester moieties that function as host recognition cues or sex pheromones, however, their specificity for these compounds remains unclear. Here, we evaluate expression levels of this gene family in the light-brown apple moth, Epiphyas postvittana, via RNAseq and identify putative odorant degrading enzymes. We then solve the apo-structure for EposCCE24 by X-ray crystallography to a resolution of 2.43 Å and infer substrate specificity based on structural characteristics of the enzyme's binding pocket. The specificity of EposCCE24 was validated by testing its ability to degrade biologically relevant and non-relevant sex pheromone components and plant volatiles using GC–MS. We found that EposCCE24 is neither capable of discriminating between linear acetate-ester odorant molecules of varying chain length, nor between molecules with varying double bond positions. EposCCE24 efficiently degraded both plant volatiles and sex pheromone components containing acetate-ester functional groups, confirming its role as a broadly-tuned odorant degrading enzyme in the moth olfactory organ.

Published
2023
Full Text
View/download PDF

45. Transformer-based deep learning for predicting protein properties in the life sciences

Author

Abel Chandra, Laura Tünnermann, Tommy Löfstedt, and Regina Gratz

Subjects
deep learning, transformers, life sciences, protein property prediction, machine learning, Medicine, Science, Biology (General), QH301-705.5
Abstract

Recent developments in deep learning, coupled with an increasing number of sequenced proteins, have led to a breakthrough in life science applications, in particular in protein property prediction. There is hope that deep learning can close the gap between the number of sequenced proteins and proteins with known properties based on lab experiments. Language models from the field of natural language processing have gained popularity for protein property predictions and have led to a new computational revolution in biology, where old prediction results are being improved regularly. Such models can learn useful multipurpose representations of proteins from large open repositories of protein sequences and can be used, for instance, to predict protein properties. The field of natural language processing is growing quickly because of developments in a class of models based on a particular model—the Transformer model. We review recent developments and the use of large-scale Transformer models in applications for predicting protein characteristics and how such models can be used to predict, for example, post-translational modifications. We review shortcomings of other deep learning models and explain how the Transformer models have quickly proven to be a very promising way to unravel information hidden in the sequences of amino acids.

Published
2023
Full Text
View/download PDF

46. Psychosocial School Conditions and Mental Wellbeing Among Mid-adolescents: Findings From the 2017/18 Swedish HBSC Study

Author

Sara Brolin Låftman, Bitte Modin, Maria Granvik Saminathen, Viveca Östberg, Petra Löfstedt, and Kristiina Rajaleid

Subjects
school demands, teacher support, classmate support, wellbeing, school, Public aspects of medicine, RA1-1270
Abstract

Objectives: To investigate mid-adolescent boys’ and girls’ experiences of school demands, teacher support, and classmate support, and explore the associations of these factors with mental wellbeing.Methods: Data were derived from the Swedish Health Behaviour in School-aged Children (HBSC) study of 2017/18, with information collected among 1,418 students in grade 9 (∼15–16 years). School demands, teacher support, and classmate support were measured by indices based on three items each. Mental wellbeing was measured by the Short Warwick-Edinburgh Mental Wellbeing Scale (SWEMWBS). Linear regression analyses were performed.Results: Higher demands were associated with lower mental wellbeing. Conversely, mental wellbeing increased with greater teacher support and classmate support. Interactions between demands and the support variables showed that at the lowest levels of teacher and of classmate support, mental wellbeing was low and not associated with school demands. With increasing levels of teacher and classmate support, the overall level of mental wellbeing increased and revealed an inverse association between school demands and mental wellbeing.Conclusion: The study contributes with knowledge about how psychosocial conditions in school may hinder or enhance wellbeing among students.

Published
2023
Full Text
View/download PDF

48. Does the family affluence scale reflect actual parental earned income, level of education and occupational status? A validation study using register data in Sweden

Author

Maria Corell, Yun Chen, Peter Friberg, Max Petzold, and Petra Löfstedt

Subjects
Family affluence scale, Validation, Sweden, Register data, HBSC, STARS, Public aspects of medicine, RA1-1270
Abstract

Abstract Aim To examine the external validity of the Family Affluence Scale (FAS) among adolescents in Sweden by using register data for parental earned income, level of education and occupational status. Methods Data from the baseline (2015–2019) of the Study of Adolescence Resilience and Stress (STARS), comprising 2283 13-year-olds in the region of Västra Götaland, were used. The FAS III consists of six items: unshared bedroom, car ownership, computer/tablet ownership, dishwasher, number of bathrooms and number of holidays abroad. Register data regarding earned income, educational level and occupational status from Statistics Sweden (2014–2018) were linked to adolescents. In total, survey data were available for 2280 adolescents, and register data were available for 2258 mothers and 2204 fathers. Results Total parental earned income was moderately correlated with adolescents’ scoring on FAS (0.31

Published
2021
Full Text
View/download PDF

49. A highly-contiguous genome assembly of the Eurasian spruce bark beetle, Ips typographus, provides insight into a major forest pest

Author

Daniel Powell, Ewald Groβe-Wilde, Paal Krokene, Amit Roy, Amrita Chakraborty, Christer Löfstedt, Heiko Vogel, Martin N. Andersson, and Fredrik Schlyter

Subjects
Biology (General), QH301-705.5
Abstract

Daniel Powell et al. present a high-quality genome assembly of the Eurasian spruce bark beetle, Ips typographus, which is known to cause substantial damage to European forests. Their results provide an important resource for investigation of the underlying physiology of this pest species and limit future threats to European forests.

Published
2021
Full Text
View/download PDF

50. A general multiblock method for structured variable selection

Author

Löfstedt, Tommy, Hadj-Selem, Fouad, Guillemot, Vincent, Philippe, Cathy, Raymond, Nicolas, Duchesney, Edouard, Frouin, Vincent, and Tenenhaus, Arthur

Subjects
Statistics - Machine Learning
Abstract

Regularised canonical correlation analysis was recently extended to more than two sets of variables by the multiblock method Regularised generalised canonical correlation analysis (RGCCA). Further, Sparse GCCA (SGCCA) was proposed to address the issue of variable selection. However, for technical reasons, the variable selection offered by SGCCA was restricted to a covariance link between the blocks (i.e., with $\tau=1$). One of the main contributions of this paper is to go beyond the covariance link and to propose an extension of SGCCA for the full RGCCA model (i.e., with $\tau\in[0, 1]$). In addition, we propose an extension of SGCCA that exploits structural relationships between variables within blocks. Specifically, we propose an algorithm that allows structured and sparsity-inducing penalties to be included in the RGCCA optimisation problem. The proposed multiblock method is illustrated on a real three-block high-grade glioma data set, where the aim is to predict the location of the brain tumours, and on a simulated data set, where the aim is to illustrate the method's ability to reconstruct the true underlying weight vectors., Comment: 24 pages, 5 figures

Published
2016

Catalog

Books, media, physical & digital resources
See catalog results