aGrUM stands for a Graphical Universal Model. It is a C++ library designed to easily implement decision-theoretic graphical models such as:

• Bayesian networks
• Influence diagrams
• Decision trees
• GAI networks
• Markov Decision processes

The library contains basic building blocks to easily implement usual graphical model tasks such as:

• Structure and parameter learning
• Elicitation of preferences
• Inference (MAP, MPE, propagation of evidence)
• Planification

The official forge website of the library is here. In particular, the website contains its repository, the reference manual and some explanations on how to use the library.

aGrUM is packaged with a Python wrapper called pyAgrum