Projects



KartAI: Smarter Property Data with Artificial Intelligence

Public sector innovation project, 2021–2027
Norwegian Research Council funding: 7 MNOK (Total budget: 15 MNOK)
Partners: Kristiansand Municipality (project owner), Norkart, Norwegian Mapping Authority, University of Agder

KartAI aims to automate and improve municipal case processing related to property registration and building permits using artificial intelligence and advanced data-driven methods. By combining AI with proactive citizen dialogue, the project will enhance efficiency, reduce manual work, and improve decision-making. With standardized data protocols and cutting-edge multimodal machine learning models, KartAI is designed for national scalability, building on the success of the earlier regional initiative KartAI 1.0. The project addresses a shared challenge across all Norwegian municipalities and seeks to set a national standard for AI-assisted public services.


Ny prosess for kunnskapsbasert vurdering og rådgiving ved bruk av kunstig intelligens ved Folkehelseinstituttet

Public sector innovation project, 2024–2027
Norwegian Research Council project number: 349547
External funding: 6.6 MNOK (Total 12 MNOK)
Partners: Norwegian Institute of Public Health (FHI), UiA, UiB

FHI, in collaboration with the University of Agder, SNF, and Kristiansand Municipality, is developing a new AI-powered process for faster and more efficient evidence-based health assessments and advice. The goal is to meet increasing demands from the health sector and general public by automating key steps in the assessment workflow. The project combines expertise in public health, artificial intelligence, and innovation management, and aims to support better decision-making both in daily operations and during crises. The innovation will also serve as a model for broader public sector transformation.


Auto3DMap: Automated 3D Mapping from Aerial Images Using AI

Public sector PhD project, 2022–2026
Norwegian Research Council project number: 338454
External funding: 2.3 MNOK (Total 4MNOK)
Lead institution: Norwegian Mapping Authority (Statens kartverk)

This project aims to automate the creation of 3D building models for national map databases using artificial intelligence and aerial imagery. By replacing manual processes with AI-driven techniques, the project will streamline the generation of building objects for the national FKB-Bygning database. Using overlapping vertical and oblique aerial images, combined with advanced deep learning and photogrammetry, the goal is to accurately reconstruct and geolocate 3D structures. The work will use real-world data from across Norway and contribute to a more efficient and scalable mapping infrastructure.


NAIC: Norwegian Artificial Intelligence Cloud

National research infrastructure project, 2023–2029
Norwegian Research Council project number: 322336
External funding: 49 MNOK
Lead institution: University of Oslo, UiA partner

NAIC aims to build Norway’s most powerful national infrastructure for artificial intelligence. The project will provide computing power, storage, and expert support to researchers across the country. It will be developed in three phases: mapping and optimizing current resources, implementing new technology and training, and finally establishing a central high-performance AI infrastructure. The platform will offer training, helpdesk services, and federated access to partner institutions. After 2029, the infrastructure will transition to national operations under Sigma2.



AUTOCASE: Autonom Saksbehandling med Maskinlæring

NæringsPhD

Norwegian Research Council project number: 346312

Partners: Norkart, UiA

AUTOCASE aims to automate municipal case processing—such as building permits and property-related decisions—using machine learning. The project focuses on developing AI models that can handle complex inputs like text, maps, and technical drawings. With over 90,000 building cases processed annually in Norway, automation offers major efficiency gains. A key part of the project is adapting public data for AI use. The project is supported by the Research Council of Norway and carried out as an industrial PhD.




AIM: AI-drevet Multiformatunivers

Innovation project in the industrial sector, 2022-2026

Norwegian Research Council project number: 332274

External funding: 13 MNOK (Total budget 30 MNOK)

Partners: Dyreparken, InFuture, Qvisten, University of Agder


Dyreparken, Qvisten Animation, inFuture, and the University of Agder are partnering on a 3-year research project called AIM, which aims to use artificial intelligence and machine learning to better understand the audience and the interaction between different formats in creating multiformat universes (large stories that unfold across different formats and evolve over time). The project is supported by The Research Council of Norway and will be collecting, combining, and analyzing data across new sources not currently used or combined by separate industries.


AI4LocalNews: Leveraging AI for Local Democracy

Industrial PhD project, 2024–2027
Norwegian Research Council project number: 354195
Partners: iTromsø, University of Agder, OsloMet

AI4LocalNews aims to develop AI tools that support local journalism by automating data analysis and enhancing newsroom efficiency. The project focuses on helping journalists uncover and communicate important local issues, while maintaining human oversight and ethical standards. Led by iTromsø, the project is carried out as an industrial PhD with support from the Research Council of Norway.


CreateView – Insight to more than what you see

Role: Workpackage leader, technical leader, and main research partner

Norwegian Research Council project number: 325862

Main funding partner: Norwegian Research Council. User-driven research-based

Innovation (BIA)

Project partners: Veterinærinstituttet, Havforskningsinstituttet, University of Agder.

Create Viewaims to develop automated tools for monitoring the health and welfare of farmed fish. The core innovation is a dead fish image scanner (CView LiftUp) and live fish monitoring (CView Eye), combined with diagnostic, production, and environmental data in a large-scale database. Using artificial intelligence and machine learning, the system analyzes patterns that are difficult for humans to detect in time. The result is faster, data-driven decision support that helps prevent fish health issues, reduces losses, improves fish welfare, and increases profitability in aquaculture.


Computer vision to expand monitoring and accelerate assessment of coastal fish 

MARINFORSK project, 2021-2025

Norwegian Research Council project number: 325862

External funding 12.2 MNOK (total budget: 13.6 MNOK)

Partners: Havforskningsinstituttet, University of Agder, University of California Santa Cruz, University of Trento, Virginia Polytechnic Institute and State University


CoastVision is a project that aims to use artificial intelligence to improve the identification and re-identification of fish in their natural habitat. They will focus on Atlantic cod, ballan wrasse and corkwing wrasse and will be the final step in a fully automated video analysis pipeline that will identify, track, size and count fish in video feeds from long-term monitoring stations.


 

Your green, smart and endless wardrobe

Innovation project in the industrial sector, 2020-2024

Norwegian Research Council project number: 309977

External funding: 23.6 MNOK

Project partners: Fjong, BI, NMBU and UIA


FJONG is a company that aims to revolutionize the way we consume clothes by making rentals more attractive than buying new ones. They are developing a unique rental platform that acts as an AirBnB for clothes, where customers can both rent outfits and lend out their own for a cut of the rental price. This project, with UiA, NBMU and BI,  expand and conduct further research to overcome key challenges such as behavioral change, sustainability and technical challenges. Behavioral change challenges include how to change consumer habits, sustainability challenges include optimizing the business model and minimizing environmental footprints, and technical challenges include utilizing artificial intelligence and machine learning to create a superior user experience and building a sharing platform that meets the requirements of the value propositions designed and the complex logistics of a peer-to-peer sharing platform. Key to the development is new artificial intelligence techniques for recommendation systems.



 

Kornmo – production optimization, quality control and sustainability through

the grain value chain

Role: Workpackage leader and main research partner

Main funding partner: Norwegian Research Council. User-driven Research based

Innovation (BIA)

Project partners: Felleskjøpet Agri SA, inFuture AS, NHO Mat og Drikke

The Kornmo-project will develop models to optimize volume, quality, and sustainability

in grain production and track the grain "from soil to table". The project will develop

machine learning approaches for the entire value chain from each shift, via grain

reception and further plant structure and end products.


 

Deep learning for sleep, mental, and emotional stage recognition from non-contact sensors of mental health patients.

Industrial PhD project, 2020–2024
Focus: Health technology, deep learning, and human activity recognition

This project developed advanced deep learning methods for monitoring human sleep and activity using non-invasive sensors. The research focused on two key areas: automatic sleep stage classification and human activity recognition (HAR), aiming to improve healthcare diagnostics and personalized monitoring.

Using EEG/EOG signals, models like TDConvNet and SleepXAI achieved near-expert accuracy in classifying sleep stages, with an emphasis on explainability for clinical use. In HAR, the project introduced privacy-preserving, interpretable video-based frameworks for fall detection and activity tracking—vital for elderly care and rehabilitation. The results contribute to more transparent, effective AI tools for healthcare, with potential for real-world integration in clinical and home settings.

 

Wonderful new world? How does the use of artificial intelligence in the customer relationship affect Sparebanken Sør's ability to act as an ethical,

socially responsible relationship player?

Role: Main research partner

Main funding partner: Norwegian Research Council. Industrial Ph.D.

Project partners: Sparebank Sør, University of Agder

The project aims to provide increased knowledge about how the use of artificial

intelligence can affect the bank's ability to act as an ethical, socially responsible

relationship player. The ambition is to provide insight that will strengthen the bank's

ability to make the right decisions in its work with digitization, automation, and the use

of artificial intelligence.


Predictive maintenance of drilling equipment using deep neural networks

Role: Main research partner

Main funding partner: Norwegian Research Council. Industrial Ph.D.

Project partners: MhWirth, University of Agder

This project explores the potential for using deep learning to predict drilling

equipment problems and estimate the remaining useful life. The project will specifically

address this problem by using real-time data from 11 rigs in operation. The focus will

be to develop and apply deep neural networks to enable autonomous decision-making

for improving service performance through condition-monitoring, predictive

maintenance, and spares management.


 

Deep Neural Networks for Energy Efficiency in High-Tech Warehouses

Role: Main research partner

Main funding partner: Norwegian Research Council. Industrial Ph.D.

Project partners: Login Eiendom, Rema 1000, University of Agder

The goal of this project is to promote the latest innovations in artificial intelligence (AI)

for energy-related operations in warehouses. More specifically, the output of this

research will be to automatize the energy-related management in a technologically

advanced warehouse to reduce the operational cost and to improve the energy

efficiency based on various data inputs and the-state-of-the-art deep neural network.


Graduate students


Ph.D. Students:



Li Meng, Graduated in 2025

State representation in Reinforcement Learning

Supervisors: Paal Engelstad, Morten Goodwin Anis Yazidi

First job after graduation: PostDoc at OsloMet

Homepage


Rupsa Saha, Graduated in 2025

Principal Supervisor: Ole-Christoffer Granmo ,Second Superviros: Morten Goodwin

Towards a Relational Tsetlin Machine in Natural Language Processing

First job after graduation: Postdoc UiA

Homepage


Micheal Dutt, Graduated in 2024

Principal Supervisor: Morten Goodwin Second Superviros: Christian Omlin

Explainable Deep Learning for Human Behaviour Understanding: Sleep Monitoring, Human Activity Recognition, and Future Opportunities for Healthcare

First job after graduation: Consultant Egde


Sven Opalic, Graduated in 2023

Principal Supervisor: Mohan Lal Kolhe, Co-supervisor: Morten Goodwin. Lei Jiao

Advanced Warehouse Energy Storage System Control Using Deep Supervised and Reinforcement Learning

First job after graduation:  Fagsjef energiteknikk og innovasjon, Ph.D. i AI til energistyring, RELOG


Rohan Kumar Yadav, Graduated in 2022

Principal Supervisor: Lei Jiao, Second Supervisor: Morten Goodwin

Interpretable Architectures and Algorithms for Natural Language Processing

First job after graduation: Senior NLP Data Scientist at Kobler AS


Nils Jakob Johannesen,Graduated in 2022

Principal Supervisor: Mohan Lal Kolhe, Second Supervisor: Morten Goodwin

Machine Learning Applications for Operation and Management of Smart Distributed Electrical Energy Network

First job after graduation: Associate Professor, University of Southern Norway

Homepage


Per-Arne Andersen, Graduated in 2022

Principal Supervisor: Morten Goodwin, Second Supervisor: Ole-Christoffer Granmo

Advancements in Safe Deep Reinforcement Learning for Real-Time Strategy Games and Industry Applications

First job after graduation: Associate Professor, University of Agder

Homepage


Darshana Abeyrathna, Graduated in 2022

Principal Supervisor: Ole-Christoffer Granmo, Second Supervisor: Morten Goodwin

Novel Tsetlin Machine Mechanisms for Logic-based Regression and Classification with Support for Continuous Input, Clause Weighting, Confidence Assessment, Deterministic Learning, and Convolution

First job after graduation: Senior Researcher DNV


Jivitesh Sharma, Graduated in 2020

Principal Supervisor: Ole-Christoffer Granmo, Second Supervisor: Morten Goodwin

Advances in Deep Learning Towards FireEmergency Application: NovelArchitectures, Techniques and Applications of Neural Networks

First job after graduation: PostDoc, University of Agder


Mehdi Ben Lazreg, Graduated in 2020

Principal Supervisor: Morten Goodwin, Second Supervisor: Ole-Christoffer Granmo

A Neural Network-Based Situational Awareness Approach for Emergency Response

First job after graduation: Machine Learning Developer  Infront ASA


Ongoing:

Karl Audun Kagnes Borgersen, Clothes Rental Recommendation and Onboarding with Deep Learning

Daniel Biermann, Deep Learning for Personalized Chatbot-based Education

Jakob Michael Voigt, Enabling Personalized Education through the use of Machine Learning and Learning Analytics

Martin Holen, A reinforcement learning-based approach for autonomous vehicles in simulated and real-world environments.

Sander Jyhne, Deep learning for map segmentation

Morten Grundetjern, Synthetic personas

Henrik Brådland, Autonom saksbehandling med maskinlæring




PostDocs:

Vimala Nunavath, 2019 - 2021

Deep learning for health analysis

First job after completion: Associate Professor University of Southern Norway


Aditya Gupta, 2021 - 

A view of more than you see of aquaculture.


Rashmi Gupta, 2021-2023

Production optimization, quality management, and sustainability through the grain value chain

Firs tjob after completion: Associate Professor Kristiania University College Norway


Jeppe Have Rasmussen, 2020 - 2023

Deep learning for the biomechanics of sound production in cod.


Graduated Master Students:

(Will come)

Publications


An almost complete list of papers can be found at my Google Scholar.