Research

Stochastic analysis

I study pathwise methods for stochastic differential equations, regularization by noise and Volterra-type systems. These problems involve irregular signals and systems whose dynamics depend on their history.

• Volterra Equations Driven by Rough Signals
• Pathwise Regularisation of Singular Interacting Particle Systems and their Mean Field Limits
• Weak Existence for Degenerate Distribution Dependent SDEs with multiplicative Noise

Signature methods and sequential data

Signature methods provide feature representations for sequential data, including time series, data streams and sequences of decisions. I also study extensions to images. My work considers the mathematical structure of these representations and their use in learning from data with memory.

• The Volterra signature
• Computational aspects of the Volterra Signature
• On the signature of an image

The tensordev repository, by Paul Hager, Luca Pelizzari and collaborators, provides computation of Volterra signatures and related objects from the preprints.

Applications and collaborations

My applied research concerns artificial intelligence, data science and modelling for retail, transportation and finance, where uncertainty, dynamics and operational decisions need to be considered together.

I lead AMOR — Center for Applied Mathematics and Operations Research at BI Norwegian Business School. The centre connects applied mathematics and operations research with industry.

I also lead SmartMatflyt, a research project with Steffen Grønneberg, Kim Van Oorschot and two PhD students. We combine mathematical modelling, machine learning and operations management to develop methods for reducing food waste and greenhouse gas emissions in the grocery sector.

I am Director of Research in SURE-AI — the Norwegian Centre for Sustainable, Risk-averse and Ethical AI. Work with the centre concerns sustainable, risk-aware and ethically grounded AI, including reliable data-driven decisions.

I am based at BI Norwegian Business School and Simula in Oslo. For research and collaboration enquiries, email fabian.a.harang@bi.no with a short description of the topic and relevant timelines.