Department of Holistic Systems

Today, everyone is talking about big-data analytics, video streaming, machine learning, web search, mobile apps, etc., and we all take it for granted that these "services" work perfectly. However, underneath such applications there are large systems, and individual components of these systems are often “only” a part of a big ecosystem of integrated building blocks needed to enable a functional service. In this respect, the HOST department addresses challenges potentially covering all components of the entire system, from data creation to visualisation of the results.
Using a combination of basic and applied research, we are targeting numerous applications, many in the field of sports and medicine, where we have a holistic view of the system and perform basic research, do experimental prototyping and run experiments in the intended environments.
Currently, machine learning is an important component of our research where we for example, in real- time, aim to analyse athlete performance and detect diseases in medical videos. In our holistic systems view, not only the accuracy of the machine learning analysis is of importance, but also the complete pipe-line integration and the system performance (e.g., resource consumption and scalability). We architect complete systems and optimise for particular application requirements, both functional and non-functional, in order to provide a best possible quality of the service and a lowest possible resource consumption. Finally, we have a goal to put the results into use for the society where we contribute to open source projects and have spun of new industries
People at Department of Holistic Systems
Who we are?
Simula Metropolitan employees are researchers, postdoctoral fellows, PhD students, engineers and administrative people. We are from all over the world, ranging from newly educated to experienced researchers, all working on making research in digital engineering at the highest international level possible.
Projects at Department of Holistic Systems
Interview training of child-welfare and law-enforcement professionals interviewing maltreated children supported via artificial avatars

The interdisciplinary FRIPRO project aims to improve interviews with maltreated children through a training program using realistic and interactive child avatars.
The department of Holistic Systems (HOST) at SimulaMet will be working with the Faculty of Social Sciences at OsloMet. The project will begin on the 1st of April 2021 and end on the 31st of March in 2024. It is funded by The Research Council of Norway with 12 million NOK and will include three Ph.D. positions.
Maltreatment and abuse of children is a significant societal problem that has serious and damaging effects on children’s behavior, psychological development, and adjustment. Detection and prevention of violence and sexual abuse against children is, therefore, a high priority for Child Protective Services (CPS) and law-enforcement professionals. The conversations and investigative interviews that are conducted with these children must be of high quality. However, both Norwegian and international research shows that despite investments in methodology, the current interview and conversation skills still need to be improved.
By using an empirically informed training system in highly realistic child avatars, this project aims to develop and maintain the advanced skills needed for interviewing maltreated children. They will use data from past investigative interviews with maltreated children and create a real-looking avatar that is capable of expressing emotion and spontaneous responses.
The planned avatar will be a combination of technologies from multiple areas in computer science including AI, computer vision, and natural language processing. The aim is for the child avatars to be a part of an interview-training program that will be implemented in cooperation with the CPS and the police. The training system will be evaluated by the project scientists to judge effectiveness in relation to real-world needs.
The project also involves collaborations with researchers from Griffith University in Australia and the University of Cambridge in the United Kingdom.
FFC: Female Football Centre

The Department of Holistic Systems at SimulaMet is collaborating with UiT - The Arctic University of Norway and Forzasys AS on the Female Football Centre (FFC), funded by the Tromsø Research Foundation.
The main goal of the centre is to gain new and fundamental insights into what affects the performance and overall health of female elite football players. A general objective is to devise novel methodologies for epidemiological research that might impact research fields in both sports and medicine. In particular, we aim to develop a non-invasive, privacy-preserving technology that enables us to continuously quantify and monitor athlete behavior where we derive analytic insights from different perspectives (e.g., biomechanics, sports-specific science, medicine, coaches, and athletes).
In the current gold standards for epidemiology, observational prospective cohort studies include that cohort subjects are followed in detail over a longer period. This is an error-prone and tedious task that has for a long time been carried out using pen and paper, and later doing a manual, tedious analysis. Making this entire process easier is the main responsibility of the researchers from the Department of Holistic Systems at SimulaMet.
In cooperation with UiT and Forzasys AS, SimulaMet has earlier developed and used an automatic performance monitoring system for athletes used by both national and elite series soccer teams. The goal is to quantify and develop accurate analysis technologies that enable a personalized assessment and performance development of elite athletes.
The automatic performance monitoring system collects athletes’ subjective parameters, like training load, wellness, injury, and illness, using a small questionnaire-app running on their mobile phones, and the data is transferred to a cloud-based backend system. Then, from a trainer-portal, the data can be automatically visualized for both individual players and team overviews.
In FFC, the objective is to extend the system further to include female-specific parameters and introduce more automatic analysis using, for example, machine learning. We will host and develop the system for all the teams participating in the project, and we will initiate automatic analyses that might be able to predict future overuse injuries or to help maybe to find the best development process of a player or a team.
Svein Arne Pettersen (head of research at the School of sports sciences, UiT) is the centre leader, and he has collaborated with the researchers at SimulaMet for a long time.
Read more about the new Female Football Centre (FFC)
FFC is funded by Tromsø Research Foundation.
Publications at Department of Holistic Systems
Journal Article
Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
Diagnostics 13, no. 14 (2023).Status: Published
Usefulness of Heat Map Explanations for Deep-Learning-Based Electrocardiogram Analysis
Deep neural networks are complex machine learning models that have shown promising results in analyzing high-dimensional data such as those collected from medical examinations. Such models have the potential to provide fast and accurate medical diagnoses. However, the high complexity makes deep neural networks and their predictions difficult to understand. Providing model explanations can be a way of increasing the understanding of “black box” models and building trust. In this work, we applied transfer learning to develop a deep neural network to predict sex from electrocardiograms. Using the visual explanation method Grad-CAM, heat maps were generated from the model in order to understand how it makes predictions. To evaluate the usefulness of the heat maps and determine if the heat maps identified electrocardiogram features that could be recognized to discriminate sex, medical doctors provided feedback. Based on the feedback, we concluded that, in our setting, this mode of explainable artificial intelligence does not provide meaningful information to medical doctors and is not useful in the clinic. Our results indicate that improved explanation techniques that are tailored to medical data should be developed before deep neural networks can be applied in the clinic for diagnostic purposes.
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Journal Article |
Year of Publication | 2023 |
Journal | Diagnostics |
Volume | 13 |
Issue | 14 |
Number | 2345 |
Date Published | 07/2023 |
Publisher | MDPI |
Keywords | electrocardiograms, Explainable artificial intelligence, heat maps |
DOI | 10.3390/diagnostics13142345 |
Enhancing Questioning Skills through Child Avatar Chatbot Training with Feedback
Frontiers in Psychology (2023).Status: Published
Enhancing Questioning Skills through Child Avatar Chatbot Training with Feedback
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Journal Article |
Year of Publication | 2023 |
Journal | Frontiers in Psychology |
Publisher | Frontiers |
DOI | 10.3389/fpsyg.2023.1198235 |
Approximate Bayesian Inference Based on Expected Evaluation
Bayesian Analysis 1, no. 1 (2023).Status: Published
Approximate Bayesian Inference Based on Expected Evaluation
Approximate Bayesian computing (ABC) and Bayesian Synthetic likelihood (BSL) are two popular families of methods to evaluate the posterior distribution when the likelihood function is not available or tractable. For existing variants of ABC and BSL, the focus is usually first put on the simulation algorithm, and after that the form of the resulting approximate posterior distribution comes as a consequence of the algorithm. In this paper we turn this around and firstly define a reasonable approximate posterior distribution by studying the distributional properties of the expected discrepancy, or more generally an expected evaluation, with respect to generated samples from the model. The resulting approximate posterior distribution will be on a simple and interpretable form compared to ABC and BSL.
Secondly a Markov chain Monte Carlo (MCMC) algorithm is developed to simulate from the resulting approximate posterior distribution. The algorithm was evaluated on a synthetic data example and on the Stepping Stone population genetics model, demonstrating that the proposed scheme has real world applicability. The algorithm demonstrates competitive results with the BSL and sequential Monte Carlo ABC algorithms, but is outperformed by the ABC MCMC.
Afilliation | Communication Systems, Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Journal Article |
Year of Publication | 2023 |
Journal | Bayesian Analysis |
Volume | 1 |
Issue | 1 |
Date Published | Jan-01-2023 |
Publisher | Project euclid |
URL | https://projecteuclid.org/journals/bayesian-analysis/volume--1/issue--1/... |
DOI | 10.1214/23-BA1368 |
Training Performance Indications for Amateur Athletes Based on Nutrition and Activity Lifelogs
Algorithms, no. 1 (2023): 30.Status: Published
Training Performance Indications for Amateur Athletes Based on Nutrition and Activity Lifelogs
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Journal Article |
Year of Publication | 2023 |
Journal | Algorithms |
Issue | 1 |
Pagination | 30 |
Date Published | Jan-01-2023 |
Publisher | MDPI |
URL | https://www.mdpi.com/1999-4893/16/1/30 |
DOI | 10.3390/a16010030 |
A multi-center polyp detection and segmentation dataset for generalisability assessment
Nature Scientific Data 10 (2023).Status: Published
A multi-center polyp detection and segmentation dataset for generalisability assessment
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Journal Article |
Year of Publication | 2023 |
Journal | Nature Scientific Data |
Volume | 10 |
Publisher | Nature |
URL | https://doi.org/10.1038/s41597-023-01981-y |
DOI | 10.1038/s41597-023-01981-y |
Live Streaming Technology and Online Child Sexual Exploitation and Abuse - A Scoping Review
Trauma, Violence, & Abuse (2023).Status: Accepted
Live Streaming Technology and Online Child Sexual Exploitation and Abuse - A Scoping Review
Livestreaming of child sexual abuse is an established form of online child sexual exploitation
and abuse. However, only a limited body of research has examined this issue. The Covid-19
pandemic has accelerated internet use and user knowledge of livestreaming services
emphasising the importance of understanding this crime. In this scoping review, existing
literature was brought together through an iterative search of eight databases containing peer-
reviewed journal articles, as well as grey literature. Records were eligible for inclusion if the
primary focus was on livestream technology and online child sexual exploitation and abuse,
the child being defined as eighteen years or younger. Fourteen of the 2,218 records were
selected. The data were charted and divided into four categories: victims, offenders,
legislation, and technology. Limited research, differences in terminology, study design, and
population inclusion criteria present a challenge to drawing general conclusions on the
current state of livestreaming of child sexual abuse. The records show that victims are
predominantly female. The average livestream offender was found to be older than the
average online child sexual abuse offender. Therefore, it is unclear whether the findings are
representative of the global population of livestream offenders. Furthermore, there appears to
be a gap in what the records show on platforms and payment services used and current digital
trends. The lack of a legal definition and privacy considerations pose a challenge to
investigation, detection, and prosecution. The available data allow some insights into a
potentially much larger issue.
Afilliation | Communication Systems, Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Journal Article |
Year of Publication | 2023 |
Journal | Trauma, Violence, & Abuse |
Publisher | SAGE Publications |
Poster
Concept Explanations for Deep Learning-Based Diabetic Retinopathy Diagnosis
Nordic AI Meet 2023, 2023.Status: Accepted
Concept Explanations for Deep Learning-Based Diabetic Retinopathy Diagnosis
Diabetic retinopathy (DR) is a common complication of diabetes that damages the eye and potentially leads to blindness. The severity and treatment choice of DR depends on the presence of medical findings in fundus images. Much work has been done in developing complex machine learning (ML) models to automatically diagnose DR from fundus images. However, their high level of complexity increases the demand for techniques improving human understanding of the ML models. Explainable artificial intelligence (XAI) methods can detect weaknesses in ML models and increase trust among end users. In the medical field, it is crucial to explain ML models in order to apply them in the clinic. While a plethora of XAI methods exists, heatmaps are typically applied for explaining ML models for DR diagnosis. Heatmaps highlight image areas that are regarded as important for the model when making a prediction. Even though heatmaps are popular, they can be less appropriate in the medical field. Testing with Concept Activation Vectors (TCAV), providing explanations based on human-friendly concepts, can be a more suitable alternative for explaining models for DR diagnosis, but it has not been thoroughly investigated for DR models. We develop a deep neural network for diagnosing DR from fundus images and apply TCAV for explaining the resulting model. Concept generation with and without masking is compared. Based on diagnostic criteria for DR, we evaluate the model’s concept ranking for different severity levels of DR. TCAV can explain individual images to gain insight into a specific case, or an entire class to evaluate overall consistency with diagnostic standards. The most important concepts for the DR model agree with diagnostic criteria for DR. No large differences are detected between the two concept generation approaches. TCAV is a flexible explanation method where human-friendly concepts provide insights and trust in ML models for medical image analyses, and it shows promising results for DR grading.
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Poster |
Year of Publication | 2023 |
Place Published | Nordic AI Meet 2023 |
Keywords | concept-based explanations, diabetic retinopathy, Explainable artificial intelligence |
Proceedings, refereed
Looking into Concept Explanation Methods for Diabetic Retinopathy Classification
In Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2023, 2023.Status: Accepted
Looking into Concept Explanation Methods for Diabetic Retinopathy Classification
Diabetic retinopathy is a common complication of diabetes, and monitoring the progression of retinal abnormalities using fundus imaging is crucial. Because the images must be interpreted by a medical expert, it is infeasible to screen all individuals with diabetes for diabetic retinopathy. Deep learning has shown impressive results for automatic analysis and grading of fundus images. One drawback is, however, the lack of interpretability, which hampers the implementation of such systems in the clinic. Explainable artificial intelligence methods can be applied to explain the deep neural networks. Explanations based on concepts have shown to be intuitive for humans to understand, but have not yet been explored in detail for diabetic retinopathy grading. This work investigates and compares two concept-based explanation techniques for explaining deep neural networks developed for automatic diagnosis of diabetic retinopathy: Quantitative Testing with Concept Activation Vectors and Concept Bottleneck Models. We found that both methods have strengths and weaknesses, and choice of method should take the available data and the end user’s preferences into account.
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Proceedings, refereed |
Year of Publication | 2023 |
Conference Name | Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2023 |
Keywords | concept-based explanations, diabetic retinopathy, Explainable artificial intelligence |
Principal Components Analysis Based Imputation for Logistic Regression
In International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems. {Springer, 2023.Status: Published
Principal Components Analysis Based Imputation for Logistic Regression
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Proceedings, refereed |
Year of Publication | 2023 |
Conference Name | International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems |
Pagination | 28–36 |
Publisher | {Springer |
Man vs. AI: An in silico study of polyp detection performance
In 2023 IEEE 36th International Symposium on Computer-Based Medical Systems (CBMS). L'Aquila, Italy: IEEE, 2023.Status: Published
Man vs. AI: An in silico study of polyp detection performance
AI-based colon polyp detection systems have received much attention, and several products and prototypes report good results. In silico verification is a crucial step when developing such systems, but very few compare human versus AI performance. This paper, therefore, describes methods and results for an in silico test of an AI model with two different versions for polyp detection in colonoscopy and compares them to the performance of endoscopist doctors who reviewed the same colonoscopy video clips. The two versions have different thresholds for false positive rate reduction. Our models perform polyp detection within the range of the endoscopists' performance, although faster, showing a potential for use in a clinical setting. For the AI and the endoscopists alike, the results show a trade-off between high sensitivity and high specificity; to achieve perfect detection, one will also get abundance of false positives. This can cause alarm fatigue in a clinical setting.
Afilliation | Machine Learning |
Project(s) | Department of Holistic Systems |
Publication Type | Proceedings, refereed |
Year of Publication | 2023 |
Conference Name | 2023 IEEE 36th International Symposium on Computer-Based Medical Systems (CBMS) |
Publisher | IEEE |
Place Published | L'Aquila, Italy |
URL | https://ieeexplore.ieee.org/document/10178833 |
DOI | 10.1109/CBMS58004.2023.00307 |