Embedded Files
Thinkover at the University of Liverpool Institute for Risk and Uncertainty
Precise versus Bounded Probability
19 August 2020, Wednesday
2 pm BST, 3 pm CEST
Please request the Zoom link from A.M.Jones@liverpool.ac.uk
What does it mean to quantify epistemic uncertainty by probability? Can there be more than one way to do it? In this discussion seminar I will compare precise and bounded probability as measures for epistemic uncertainty. I look for criteria for suitable quantitative measures of epistemic uncertainty such as a coherent theory, interpretation, the ability to learn from data in different types of situations, the ability to integrate expert knowledge in a transparent way and the ability to propagate epistemic uncertainty in a model. We focus the discussion to Bayesian inference using precise subjective probability, robust Bayes inference using bounded probability, and (to some extent) confidence boxes using the theory of confidence structures. A set of challenge problems have been specified to aid the discussion.
To get the best out of the seminar you can prepare by 1) Looking at the challenge problems on the webpage and think how uncertainty can be quantified by probability if these were a real assessment and 2) Providing at least one argument for and against using probability to quantify epistemic uncertainty (you will be asked to share these during the seminar)
Note that the seminar including the discussions will be recorded and posted on the website afterwards.
Organiser Ullrika Sahlin is an associate professor in Environmental Science at Lund University, Sweden. She is doing research on the use of methods to treat uncertainty in scientific assessments with applications on environmental assessments and evidence-based decision making. Ullrika is leading the research group Uncertainty and Evidence Lab at Lund University.
Page updated
Google Sites
Report abuse