LAZY propagation: a junction tree inference algorithm based on lazy evaluation
Artificial Intelligence
Probabilistic Reasoning in Multi-Agent Systems: A Graphical Models Approach
Probabilistic Reasoning in Multi-Agent Systems: A Graphical Models Approach
Agent-encapsulated bayesian networks
Agent-encapsulated bayesian networks
Belief Update in Bayesian Networks Using Uncertain Evidence
ICTAI '06 Proceedings of the 18th IEEE International Conference on Tools with Artificial Intelligence
Semantically-linked bayesian network: a framework for probabilistic inference over multiple bayesian networks
An approach to hybrid probabilistic models
International Journal of Approximate Reasoning
On the revision of probabilistic beliefs using uncertain evidence
Artificial Intelligence
Hi-index | 0.00 |
The goal of this thesis is to allow easier design of probability-based agents and multiagent systems, resulting in rational decision making. A multiagent framework is presented and compared with other proposed frameworks where advantages and disadvantages of each are outlined. A central problem of message passing in probabilistic systems is the familiar rumor problem, where cycles in message passing cause redundant influence of beliefs. We develop algorithms to identify and solve the rumor problem in the context of our multiagent system. Central to our message passing scheme is the notion of soft evidential update. Traditional propagation algorithms are not compatible with soft evidence. We propose a new propagation algorithm that is based on Lazy propagation and compare the theoretical and experimental performance with other proposed solutions.