Subjective Logic
by Audun Jøsang Tutorial on subjective logic given at UAI 2016 in New York.Belief Reasoning with Subjective Logic
Subjective logic is a belief reasoning calculus that is compatible with, and an extension of probabilistic logic. It can for example be used for modeling trust networks, for modelling Bayesian networks, for Inteligence Analysis and logical argumentation. In general, subjective logic is suitable for modeling and analysing situations involving uncertainty, incomplete knowledge and different world views.
Beliefs are represented as opinions. A binomial opinion applies to a single proposition, and a multinomial opinion applies to a whole frame of multiple propositions (state space). Binomial opinions are equivalent to Beta probability density functions, and multinomial opinions are equivalent to the more general Dirichlet probability density functions. This makes subjective logic suitable for reasoning with evidence represented as Beta or Dirichlet probability density functions. For details and theory see the the new book on subjective logic or the Wikipedia subjective logic page.
Subjective Logic Demonstrations
- See interpretations of binomial opinions with the belief visualisation demo.
- Play with subjective logic operators.
- Analyse a simple trust network with subjective logic.
- Excel spreadsheet for 3x3 conditional deduction and abduction
Back to Homepage.
Last Updated 24 March 2011