01
Signature methods in Data Science
Feature representations for paths, images and time series, with an emphasis on memory, structure and interpretable learning.
About
Research with memory, uncertainty and structure.
I develop data science methods for systems that are irregular, history-dependent or high-dimensional, with applications in industry, public-facing advice and decision-making under uncertainty.
My research sits at the intersection of signature methods in data science, stochastic analysis, applied stochastic modelling and trustworthy AI. A common thread is the search for precise mathematical representations of complex dynamics, especially when memory, uncertainty and irregularity are central rather than peripheral.
I am Professor of Data Science at BI Norwegian Business School and Adjunct Chief Research Scientist at Simula Research Laboratory. This work connects naturally with data science, operations research and trustworthy AI: designing methods and advice that are not only powerful, but interpretable, reliable and useful for real decision environments.
The ambition is simple: make complex systems more understandable, more useful and more responsible.
Research
Current research interests
Four themes currently shape my research, ranging from mathematical foundations to industry-facing modelling and AI decision support.
02
Stochastic Analysis
Pathwise methods, stochastic differential equations, regularization by noise and Volterra-type systems.
03
Applied Stochastic modelling with applications in Industry
Stochastic models for industry-facing problems where uncertainty, dynamics and operational decisions need to be handled together.
04
Trustworthy AI and Decisions
Reliable data-driven tools and thoughtful AI advice for decisions under uncertainty, developed with AMOR, SURE-AI and collaborators.
Software for Volterra signatures
The tensordev repository by Paul Hager, Luca Pelizzari and collaborators provides efficient computation of Volterra signatures and related objects from recent preprints.
AI Notes
Thoughts on AI
A dedicated place for notes, talks and public-facing reflections on trustworthy AI, mathematical risk and responsible data-driven decisions.
On creating a strategic Norwegian-European AI fund
Together with Andreas Ravndal Kostøl, I have written some thoughts on developing a strategic Norwegian-European AI investment fund constructed from the Norwegian Pension Fund and built around the same model, with the objective of strengthening European AI infrastructure and industry.
Read more on practical implementationSustainable, risk-averse and ethical AI
Research and collaboration on AI systems that are sustainable, risk-aware and ethically grounded, connected to the SURE-AI centre.
Visit SURE-AINews
Updates from AMOR and SURE-AI
A simple news area for selected AMOR and SURE-AI updates, AI notes, notes, talks, reports and other items worth highlighting.
AMOR updates
Selected centre news, activities and research highlights can be posted here.
AMOR centreSURE-AI updates
Relevant news from SURE-AI, especially around trustworthy AI and responsible data-driven decisions.
Visit SURE-AIThoughts on AI
Public reports, advisory documents and short reflections on AI can be collected in the AI Notes section.
AI NotesPublications
Recent and selected work
A complete, compact publication list is included below, with direct links to arXiv records where available. Google Scholar remains the best place for citation counts and profile-level updates.
2026
2026
2025
2025
2024
2024
Stochastic Processes and their Applications, 187, 104661, 2025. DOI: 10.1016/j.spa.2025.104661
2023
2023
2022
2022
2022
2021
2021
Distribution dependent SDEs driven by additive continuous noise
Electronic Journal of Probability, 27, 1-38, 2022. DOI: 10.1214/22-EJP756
2021
2021
2020
Pathwise Regularisation of Singular Interacting Particle Systems and their Mean Field Limits
DOI: 10.1016/j.spa.2023.02.005
2020
2020
2020
2020
2020
2019
2019
2018
2018
2018
2017
Centres
AMOR and SURE-AI
Research leadership and collaboration live across centres that connect mathematics, operations research and trustworthy artificial intelligence.
AMOR
Center for Applied Mathematics and Operations Research.
Contact about AMORSURE-AI
The Norwegian Centre for Sustainable, Risk-averse and Ethical AI.
Visit SURE-AIContact
Get in touch
The easiest way to reach me is by email. For research and collaboration requests, please include a short description of the topic and relevant timelines.
Email
fabian.harang@bi.no
Location
BI Norwegian Business School and Simula Research Laboratory, Oslo
AI Notes
Thoughts, notes and documents
Scholar
Google Scholar profile