Alexander Selvikvåg Lundervold

Arbeids- og kompetanseområde

Hvordan kan vi transformere kunstig intelligens-metoder til praktiske, KI-baserte løsninger? Hvilken rolle kan KI spille innen sektorer som utdanning, medisin og helse både nå og i fremtiden?

Alexander er professor i kunstig intelligens (KI) ved Høgskulen på Vestlandet (HVL) og en engasjert formidler av KI og dens praktiske anvendelser. Hans arbeid ligger i skjæringspunktet mellom maskinlæring, kunstig intelligens og dataingeniørfag, med særlig fokus på utvikling, evaluering og implementering av KI-baserte løsninger innen medisin, helse og utdanning.

Ved HVL leder han arbeidet med KI-basert innovasjon og effektivisering gjennom roller i KI-koordineringsgruppen, forskningsgruppen for Artificial Intelligence Engineering og HVLs KI-lab. Han underviser i kunstig intelligens for undervisere og administrativt ansatte, og gir jevnlig råd til virksomheter om hvordan KI kan forbedre tjenester og åpne for nye muligheter.

Siden 2018 har han ledet aktivitetene innen Medisinsk AI ved Mohn Medical Imaging and Visualization Centre (MMIV) på Haukeland universitetssjukehus, med fokus på utvikling av maskinlæringsdrevet programvare for medisinsk diagnose og behandling.

Formidling

Alexander har holdt over 100 foredrag for helsesektoren, akademia, næringsliv og offentlige organisasjoner. Foredragene spenner over fire hovedområder: 

Helse og medisin: Medisinsk KI og fremtidens helsetjenester for sykehus, helseforetak og medisinske konferanser.

Utdanning og akademia: KI i undervisning, forskningsinnovasjon og digitale læringsverktøy for universitets- og høgskolesektoren og videregående skoler. 

Næringsliv og innovasjon: Praktiske anvendelser av KI for effektivisering og forretningsutvikling, i store konsern og lokale bedriftsnettverk. 

Samfunn og etikk: Offentlige foredrag og paneldebatter om KIs rolle i samfunnsutviklingen, demokrati og etiske perspektiver. 

Prosjekt

• ASIS - AI-supported services for image-diagnostics in Western Norway, funded by the Western Norway Regional Health Authority (2025-2029)
• AIMS Norway – Artificial Intelligence in Mammography Screening in Norway, funded by the Western Norway Regional Health Authority (2023–2025).
• Part of the project leadership and PI in a machine learning work package in WIML: Workflow-integrated machine learning at the MMIV, funded by the Norwegian Research Council (2020–2024).
• PI of a work package in the project AI-Support in Medical Emergency Calls: The AISMEC-project, funded by the Norwegian Research Council (2022–2025), led by Guttorm Brattebø from Helse Bergen HF and KoKom.
• Partner in AkademiX, 2023-
• Part of the coordinating team of the Norwegian research network PRESIMAL: Precision imaging and machine learning for better patient care, funded by Nasjonal samarbeidsgruppe for helseforskning i spesialisthelsetjenesten (2021-2023), with partners from all the health regions in Norway and their universities.
• Co-PI in a machine learning work package in the Digital Life Norway project Towards better computational approaches and responsible innovation strategies in early drug discovery – application to antibiotics and COPD (2019–2023), led by Nathalie Reuter from UiB.
• Co-PI of the project Computational medical imaging and machine learning – methods, infrastructure and applications, funded by the Trond Mohn Foundation (2018–2022).
• Co-PI in a work package of the project Imaging biomarkers for precision medicine in Acute Myeloid Leukemia (AML), led by Cecilie Brekke Rygh from HVL, funded by the Western Norway Regional Health Authority (2020–2022). The main objective of the project is to evaluate the role of PET and PET-derived predictive imaging biomarkers in assessing early treatment response in AML patients to improve overall outcomes.
• Member of the project Precision imaging in gynecologic cancer at MMIV, led by prof. dr. med. Ingfrid Haldorsen. The aim of the project is to integrate imaging biomarkers into clinically relevant treatment algorithms for gynecologic cancers.
• Member of the project Disrupt, potentiate and rewire — a novel framework for understanding electroconvulsive therapy at MMIV, led by Leif Oltedal , financed by the Western Norway Regional Health Authority.
• Member of the project Deep learning in image diagnostics: transfer learning and active learning for efficient use of data and radiological expertise at MMIV, led by Sathiesh Kaliyugarasan, financed by the Western Norway Regional Health Authority (2020-2023).
• Member of the project From cognitive aging to dementia — a longitudinal imaging-based machine learning approach, led by Alexandra Vik, financed by the Western Norway Regional Health Authority (2020-2023).
• Part of Kunstig intelligens i norsk helsetjeneste (KIN), a national network for artificial intelligence in health care. I was part of the coordinating team of the network in the period 2020—2022.
• Head of the project group in an AI committee established by Helse Vest RHF. The goal is to investigate machine learning based software solutions for imaging diagnostic support that could potentially be useful in the established radiological workflow in Helse Vest.
• Member of a committee established by the Faculty of Medicine, UiB. Our report (Aug. 2020) proposed a plan for establishing Medical AI as a cross-institutional and cross-disciplinary field of research, innovation and education in Bergen.

Veiledning

PhD

• Sathiesh Kaliyugarasan: Deep learning in image diagnostics: transfer learning and active learning for efficient use of data and radiological expertise. Funded by the Western Norway Regional Health Authority (2020–2023). He defended his thesis October 3^rd, 2023
• Samaneh Abolpour Mofrad (2018–2021): Learning and Cognition in Brain and Machine: Prediction of dementia from longitudinal data and modelling memory networks. She defended her thesis November 26, 2021.

Bi-veiledning

Pågående

• Kasia Kazimierczak, New strategies for analysis of resting state fMRI, together with Karsten Specht (main supervisor) and Vince Calhoun.
• Emil Kristoffer Iversen, Artificial intelligence support in stroke calls: The AISI-study, together with Guttorm Brattebø (main supervisor), Anette Fromm and Hege Ihle-Hansen.

Fullførte

• Muhammad Ammar Malik, Unsupervised and scale-free discovery of genetic factors influencing brain structure and function, together with Tom Michoel (main supervisor) and Inge Jonassen, Department of Informatics, UiB.
   

MSc

• Eilert Skram and Daniel Kristiansen Gunleiksrud (2023–2025). AI and education: Constructing and evaluating an LLM-based course assistant.
• Øyvind Grutle and Jens Andreas Thuestad (2021–2023). Speech-to-text models to transcribe emergency calls (EMCC / 113)
• Kjetil Dyrland (2020–2022). Evaluation and Improvement of Machine Learning Algorithms in Drug Discovery.
• Jostein Digernes and Carsten Ditlev-Simonsen (2020–2022). A workflow-integrated brain tumor segmentation system based on fastai and MONAI.
• Anders Benjamin Grinde and Bendik Johansen (2019–2021). Using Natural Language Processing with Deep Learning to Explore Clinical Notes.
• Malik Aasen and Fredrik Fidjestøl Mathisen (2019–2021). De-identification of medical images using object-detection models, generative adversarial networks and perceptual loss.
• Adrian Storm-Johannessen and Sondre Fossen-Romsaas (2018–2020). Medical image synthesis using generative adversarial networks.
• Sivert Stavland (2018–2020). Machine learning and electronic health records.
• Sindre Eik de Lange and Stian Heilund (2017–2019). Autonomous mobile robots: Giving a robot the ability to interpret human movement patterns, and output a relevant response.
• Sathiesh Kumar Kaliyugarasan (2017–2019). Deep transfer learning in medical imaging. A study of how to best use transfer learning when training deep neural networks for biomedical image analysis.
• Sean Meling Murray (2017–2018). An Exploratory Analysis of Multi-Class Uncertainty Approximation in Bayesian Convolutional Neural Networks.

BSc

• Preben Andersen and Andrea M. Svendheim (2024). LLMs and fish health. In collaboration with Lerøy Seafood Group
• Harald Giskegjerde Nilsen and Sindre Kjeldrud (2024). LLMs for health advice. In collaboration with the Faculty of Medicine, UiB.
• Bendik Mathias Johansen and Kathinka Neteland (2019). Automating Reports on Water Consumption and Availability. A data science project together with Bouvet and Bergen Vann.
• Jon Einar Haraldsvik, Stian Gudvangen Gjerløw, Didrik Fanuelsen Tranvåg (2015). Tryg Maintenance App – A cross-platform application using Appcelerator Studio Cloud Services and Arrow DB. The students developed a cross-platform mobile application for Tryg Forsikring. The project was awarded "best bachelor project" at the department in 2016. The students went on to start Appivate AS.

Postdocs, main mentor

Completed

• Alexandra Vik: From cognitive aging to dementia – a longitudinal imaging-based machine learning approach. Funded by the Western Norway Regional Health Authority (2020–2022).
• Piero Mana. Worked in the RESPOND3 drug discovery research project. Funded by the Norwegian Research Council (2020–2023).

Underviser i

Jeg har opprettet flere kurs i maskinlæring og kunstig intelligens for programvareutviklere ved HVL. Disse danner basisen for spesialiseringer på bachelor- og mastergradsnivå ved HVL. Jeg underviser også maskinlæring og kunstig intelligens for medisinere, helsearbeidere og lærere.

Kurs

• DAT158: Machine learning engineering. Et praktisk, "hands-on", prosjektbasert utforsking av grunnleggende maskinlæring. Fokuserer på anvendelser av maskinlæring og hvordan prinsipper fra programvareutvikling benyttes for vellykket utvikling av maskinlærings-baserte system.
• DAT255: Deep learning engineering. Masterkurs om praktiske anvendelser av dyplæring og konstruksjon av dyplæringsbaserte applikasjoner.
• FD28: Kunstig intelligens i utdanning
• ADA524: Large language models. A comprehensive introduction to LLMs within the scope of applied computer science and engineering. Foundational theory, practical tools, and methodologies that drive LLMs' current development and application. 
• DAT801: Maskinlæring for forretningsutvikling
• ELMED219: Kunstig intelligens og beregningsorientert medisin. Et samarbeid mellom Inst. for biomedisin, Det medisinske fakultet, UiB, og Inst. for datateknologi, elektroteknologi og realfag, HVL. Kurset tilbys både medisinerstudenter og ingeniør-studenter, og forsøker å fremme økt samarbeid mellom disiplinene.
• HVL-DLN-AI: A hands-on course on artificial intelligence in computational biotechnology and medicine
• PCS956: Recent trends in applied machine learning

Forsker på 

• Maskinlæring
• Kunstig intelligens
• Dataanalyse
• Medisinsk AI
• Beregningsorientert medisin

Min forskning er knyttet til mine stillinger som professor ved Institutt for datateknologi, elektroteknologi og realfag ved HVL og "senior data scientist" ved Haukeland Universitetssjukehus. Forskningen bygger på min bakgrunn fra matematikk, mastergrad og Ph.d. i matematikk fra Universitetet i Bergen, og mitt arbeid i stilling som postdoktor ved Institutt for matematiske fag ved NTNU og som Marie Curie Fellow i gruppen Advanced Learning and Evolutionary Algorithms ved INRIA, Bordeaux (https://www.inria.fr).

Forskergrupper

• HVL Machine Learning Engineering
• MMIV-Medical AI
• HVL Data Science Group