A Complete Criterion for Value of Information in Soluble Influence Diagrams

  title={A Complete Criterion for Value of Information in Soluble Influence Diagrams},
  author={Chris van Merwijk and Ryan Carey and Tom Everitt},
Influence diagrams have recently been used to analyse the safety and fairness properties of AI systems. A key building block for this analysis is a graphical criterion for value of information (VoI). This paper establishes the first complete graphical criterion for VoI in influence diagrams with multiple decisions. Along the way, we establish two techniques for proving properties of multi-decision influence diagrams: ID homomorphisms are structure-preserving transformations of influence… 

Figures from this paper


A note about redundancy in influence diagrams
Evaluating Influence Diagrams using LIMIDs
This paper shows how to convert an influence diagram to a LIMID and describes the procedure for finding an optimal strategy, which can yield significant savings of memory and computational time when compared to traditional methods.
Submodel Decomposition Bounds for Influence Diagrams
A model decomposition framework in both IDs and LIMIDs is presented, which is called submodel decomposition that generates a tree of single-stage decision problems through a tree clustering scheme and develops a valuation algebra over the submodels that leads to a hierarchical message passing algorithm that propagates conditional expected utility functions over a submodel-tree as external messages.
Evaluating Influence Diagrams
An algorithm is developed that can evaluate any well-formed influence diagram and determine the optimal policy for its decisions and can be performed using the decision maker's perspective on the problem.
Influence Diagram Retrospective
A retrospective discusses the evolution and application of influence diagrams, which permits high-level graphic communication, clear assessments and computation in a single graphical system.
Welldefined Decision Scenarios
This paper presents a set of conditions which are necessary and sufficient to ensure that a partial influence diagram is welldefined, and uses these conditions as a basis for the construction of an algorithm for determining whether or not a partial Influence diagrams is well defined.
PyCID: A Python Library for Causal Influence Diagrams
The Python package PyCID is introduced, built upon pgmpy, that implements (causal) influence diagrams, a widely used graphical modelling framework for decision-making problems that helps answer questions about behaviour and incentives in both single-agent and multi-agent settings.
Bayes-Ball: The Rational Pastime (for Determining Irrelevance and Requisite Information in Belief Networks and Influence Diagrams)
A new, simple, and efficient "Bayes-ball" algorithm is presented which determines irrelevant sets and requisite information more efficiently than existing methods, and is linear in the size of the graph for belief networks and influence diagrams.
Agent Incentives: A Causal Perspective
A framework for analysing agent incentives using causal influence diagrams is presented and a new graphical criterion for value of control is proposed, establishing its soundness and completeness and introducing two new concepts for incentive analysis.
Decisions and Dependence in Influence Diagrams
The concept of dependence among variables in a Bayesian belief network is well understood, but what does it mean in an influence diagram where some of those variables are decisions? There are three