Explore our knowledge

Søren Wen­gel Mo­gensen

Associate Professor

Subjects
Mathematics Statistics Artificial intelligence Machine learning Data

Primary research areas

Stat­ist­ics
I de­vel­op new meth­ods in the­or­et­ic­al stat­ist­ics and ma­chine learn­ing and ap­ply them to chal­len­ging prob­lems in en­gin­eer­ing, epi­demi­ology, and eco­nom­ics.
Caus­al in­fer­ence and caus­al dis­cov­ery
Ques­tions re­lat­ing to cause and ef­fect are cent­ral in sci­ence. I work on data-driv­en meth­ods that may give us a caus­al un­der­stand­ing of real-world phe­nom­ena (caus­al in­fer­ence). I also work on meth­ods for learn­ing caus­al graphs from data (caus­al dis­cov­ery).
Graph­ic­al mod­els
Graphs are use­ful rep­res­ent­a­tions of stat­ist­ic­al mod­els. I de­vel­op meth­ods for rep­res­ent­ing stat­ist­ic­al de­pend­en­cies and caus­al re­la­tions us­ing graphs.

I study ways to go from data to know­ledge

Many questions in science, everyday life, and business are causal in nature, and understanding cause and effect is important when making decisions. If we observe a positive correlation between wine consumption and living to old age, should we then recommend drinking a glass of wine daily? Income may be a so-called confounder: Perhaps people with above-average incomes drink more wine and live longer lives which could explain the observed correlation, even if there is no direct causal link between wine consumption and longevity. In this case, increasing wine consumption may not increase life expectancy.

In society and business, understanding the effects of potential actions is crucial for informed decision making. My research focuses on statistical methods for answering causal questions, especially in systems that evolve over time such as the climate or the economy. 

Links