The half-day workshop featured a strong lineup of presentations that covered approaches to time series analysis, including generative AI, fiber optic data interpretation, predictive maintenance, and statistical methods for change detection.
Alireza Naziri Khansari opened the workshop with an overview of the evolving landscape of generative AI for time series, highlighting emerging models such as TimeGPT, TimesFM, and Chronos. His talk framed the broader discussion of how AI is reshaping the interpretation and forecasting of time-dependent data across sectors.
Silje Fuglerud Schwermer followed with insights into Equinor’s use of fiber optic monitoring for asset surveillance. She discussed the complexities of processing high-resolution, spatial-temporal data and the technical challenges of classifying and interpreting such datasets.
Equinor’s Hugo Bettencourt Machado presented Omnia.Prevent, the company’s predictive maintenance platform. Now in its fifth year of production use, the system leverages machine learning to anticipate equipment failures, supporting more efficient maintenance planning and reducing downtime.
From the University of Oslo, Arne Bang Huseby provided an academic perspective on predictive maintenance strategies, outlining the methodological foundations and real-world challenges. He emphasised the importance of condition-based monitoring as a proactive alternative to traditional maintenance models.
Concluding the session, Per August Moen from UiO delved into recent advances in change and anomaly detection. He advocated for the application of statistically grounded methods and shared case studies with direct relevance to dScience partners.
These types of events are a core feature of the dScience Partner Program, providing a unique arena for partners to share insights, showcase research, and engage in cross-sector dialogue on emerging technologies. By bringing together industry and academia in focused, thematic workshops, the program fosters collaboration, accelerates innovation, and ensures that cutting-edge research is aligned with real-world challenges and opportunities.
A special thanks to Patrick Blomquist at Equinor for organising the event, and to all dScience partners who participated and contributed to the discussions.
The next TWG workshop will be hosted by DNV, continuing the Partner Program’s mission to strengthen collaboration across the AI and data science ecosystem.
See more photos from the seminar
Photos: Christoffer Hals, dScience, UiO.