International Joint Conference on Artificial Intelligence · 2013 · 29 citations · 17 references
Artificial IntelligenceHypothesis SpaceEngineeringAgent Decision-makingIntelligent SystemsTask PlanningHeuristic Weighted ModelOperations ResearchData SciencePattern RecognitionSystems EngineeringRobot LearningWeighted ModelMulti-agent PlanningMachine VisionProbabilistic SystemModel CountingProbability TheoryComputer ScienceAi PlanningPlan Execution ModelsAutomated ReasoningHeuristic PlanningStatistical InferenceProbabilistic Programming
Understanding other agents’ goals and plans is crucial for intelligent behaviour, and weighted model counting can recognise interleaved plans but suffers from combinatorial explosion that limits real‑world applicability. The paper proposes a heuristic weighted model counting algorithm that limits generated plan execution models to quickly recognise goals by computing lower and upper bound likelihoods. Given a plan library and observations, the algorithm enumerates consistent plan execution models, computes goal probabilities as proportions of models satisfying the goal, and uses a heuristic to limit model generation while estimating lower and upper bound likelihoods.
The ability to understand the goals and plans of other agents is an important characteristic of intelligent behaviours in many contexts. One of the approaches used to endow agents with this capability is the weighted model counting approach. Given a plan library and a sequence of observations, this approach exhaustively enumerates plan execution models that are consistent with the observed behaviour. The probability that the agent might be pursuing a particular goal is then computed as a proportion of plan execution models satisfying the goal. The approach allows to recognize multiple interleaved plans, but suffers from a combinatorial explosion of plan execution models, which impedes its application to real-world domains. This paper presents a heuristic weighted model counting algorithm that limits the number of generated plan execution models in order to recognize goals quickly by computing their lower and upper bound likelihoods.
17