I am a Senior Researcher at NORCE Research AS, specializing in machine learning and its applications to infrastructure monitoring, healthcare technology, and aquaculture. My research focuses on developing AI systems that work reliably with real-world sensor data, from fiber optic sensing in pipelines to computer vision for subsea inspection. Recent directions include physics-informed methods for energy systems, large language model tooling, and international research partnerships with institutions in Europe and Southeast Asia.
Research Interests
Machine Learning · Deep Learning · Computer Vision · Distributed Fiber Optic and Seismic Sensing · Physics-Informed Neural Networks · Knowledge Distillation · Large Language Models and Tool-Calling Agents · 3D Point Cloud Processing · Healthcare Technology
Current Work
- AQUAROM: Aquaculture Disease Management — International collaboration with partners in Norway, the Philippines, and Indonesia, investigating whether smart algorithms with minimal sensors can match expensive monitoring systems for early disease detection in fish farming. The consortium spans seven institutions and includes three PhD positions across Norway and Southeast Asia, building long-term research capacity in regions that depend on aquaculture.
- NOR-DMT: Digital Music Therapy — AI-powered platform for virtual music therapy, fully tested and ready for deployment on a secure, GDPR-compliant infrastructure. The platform brings together real-time video conferencing with machine learning that analyses patient engagement, and supplies therapists with structured clinical reporting tools. Stack: Jitsi Meet (Prosody, Jicofo, JVB) for video, NestJS and PostgreSQL on the backend, React frontend over Socket.IO, Docker Compose for deployment.
- Distributed Sensing for Infrastructure and the Subsurface — Machine learning built on distributed acoustic and temperature sensing (DAS/DTS) for leak detection in water pipelines and multiphase fluid flow estimation in industrial settings. Recent work extends this to passive seismic sensing for groundwater table estimation using fiber optic instrumentation (S-TRANET). Stack: PyTorch and scikit-learn for modelling, ObsPy and SciPy for seismic and acoustic signal processing, dispersion-curve analysis with disba, HPC for training and ensemble runs.
- Computer Vision for Offshore Inspection — Real-time deep learning that detects and classifies objects in subsea imagery, from pipelines and equipment to marine life, including a live streaming inspection pipeline developed with industry partners. Stack: RT-DETR for detection, SAM-3 for weak-supervised annotation, PyTorch for training, FastAPI with Server-Sent Events for live streaming.
- Physics-Informed ML and High-Performance Computing (NAIC) — Contributions to the Norwegian Artificial Intelligence Cloud, including climate teleconnection analysis, physics-informed neural networks for energy systems such as green hydrogen electrolyzers, and hybrid optimization algorithms. Datasets and demonstrators from this work are published openly. Stack: PyTorch and scikit-learn for ML, knowledge distillation (teacher–student) for physics-informed networks, Butler-Volmer and Nernst electrochemistry for physics constraints, hybrid optimisation combining Deterministic Crowding GA with CMA-ES, NetCDF and HPC for climate ensembles.
- LLM Tooling for Data Exploration and Digital Twins — Chat-driven interfaces that connect natural language to data exploration, visualization, and digital twin systems. The models call tools from a constrained catalog and return typed, auditable outputs rather than free-form text. Stack: Models: locally hosted Mistral for text and Qwen Vision for multimodal input; Tool integration: Model Context Protocol (MCP) for tool catalogs and structured render specifications; Data and transport: NetCDF for scientific datasets and Protocol Buffers for service-to-service communication; Front end: typed React components rendered from tool outputs. Related work looks at detecting and reducing hallucinations in large language model outputs.
Education
Ph.D. in Applied Informatics, Norwegian University of Life Sciences (2020) M.Sc. in Informatics, Bandung Institute of Technology (2016) B.Sc. in Computer Science, Sepuluh Nopember Institute of Technology (2011)
Publications
- HA Arief, PJ Thomas, W Li, C Brekken, M Hjelstuen, IE Smith, S Kragset, AK Katsaggelos. Nonlinear interpolated Variational Autoencoder for generalized fluid content estimation. Geoenergy Science and Engineering, 2025
- F Hasan, H Ali, HA Arief. From Mesh to Neural Nets: A Multi-method Evaluation of Physics Informed Neural Network and Galerkin Finite Element Method for Solving Nonlinear Convection–Reaction–Diffusion Equations. International Journal of Applied and Computational Mathematics, 2025
- HA Arief, PJ Thomas, T Wiktorski. Better modeling out-of-distribution regression on distributed acoustic sensor data using anchored hidden state mixup. IEEE Transactions on Industrial Informatics, 2022
- HA Arief, PJ Thomas, K Constable, AK Katsaggelos. Towards Building a Distributed Virtual Flow Meter via Compressed Continual Learning. Sensors, 2022
- HA Arief, T Wiktorski, PJ Thomas. A survey on distributed fibre optic sensor data modelling techniques and machine learning algorithms for multiphase fluid flow estimation. Sensors, 2021
- HA Arief, M Arief, G Zhang, Z Liu, M Bhat, UG Indahl, H Tveite, D Zhao. SAnE: Smart annotation and evaluation tools for point cloud data. IEEE Access, 2020
- HA Arief, UG Indahl, GH Strand, H Tveite. Addressing Overfitting on Point Cloud Classification using Atrous XCRF. ISPRS Journal of Photogrammetry and Remote Sensing, 2019
- HA Arief, GH Strand, H Tveite, UG Indahl. Land cover segmentation of airborne LiDAR data using stochastic atrous network. Remote Sensing, 2018
Links
Google Scholar · ResearchGate · Cristin · GitHub
Last updated: May 2026 — tone and grammar edited by AI
