Statistical modelling, Machine learning and Artificial intelligence (AI)
- Department SAMBA, SAND
- Industries involved Climate and environment, Finance and insurance, Health, Natural resources, Technology and industry, Community
We develop methods and practical applications that encompass the spectrum of statistical modelling and machine learning. The very components of artificial intelligence, our expertise enables us to harness the power of AI to solve problems in a range of research areas and in multiple industries.
Statistical modelling
Statistical modelling is about understanding contexts and making forecasts based when coincidences and uncertainty are involved.
While statistical modelling and machine learning are sometimes used interchangeably, machine learning is often able to detect patterns without the use of models. As we have comprehensive experience and knowledge in both areas, we are able to customise our methods to find the most accurate solution for each problem.
Machine learning
Machine learning enables computers to automatically detect patterns in data, and the objective is to train algorithms to predict events or make decisions without being explicitly instructed to do so.
As a general tool, machine learning can be utilised in most domains, and we work with a number of clients and partners across different sectors, including healthcare, finance and climate.
We are particularly interested in enhancing transparency to better understand how systems based on machine learning systems calculate predictions and determine results. In addition to ongoing research in eXplainable artificial intelligence (XAI), we have created the digital toolkit, eXplego, which intends to ease navigation for developers working in this environment.
Language technology
Language technology is a collective term for various technologies that use natural language as data.
We work with language technology in a number of ways, including text mining, automated transcription and translation. We are especially dedicated to the development of Norwegian language technology and other models with modest amounts of training data available.
Geomodelling
We have provided advanced statistical and mathematical modelling for the oil and gas industry for almost forty years.
Geomodelling combines geological knowledge with data in a statistical model, and enables us to understand and predict natural processes on the Earth’s subsurface or geological features. Geomodels are useful for decision support in all situations where geology is a crucial factor, such as in the petroleum industry.
Research areas
Selected research topics
Anomaly detection is a research topic in growth. Anomaly detection is about observing and analysing large data sets over time, and identifying and classifying divergent behaviour.
Explainable Artificial Intelligence (XAI) provides insight into how systems based on machine learning and artificial intelligence (AI) calculate predictions and make decisions. We have created eXplego, a digital toolkit, to ease navigation for developers working in this environment.
