MSc Seminar: Abdulrahman Alshermemry

Title: Privacy Sensitive Environment Decomposition for Hypertree Agent Organization Construction by Abdulrahman Alshememry

Advisor: Dr. Yang Xiang
Advisory Committee: Dr. Charlie Obimbo

Abstract: 

Cooperative multiagent systems form an active area of research and practice in AI and software engineering. Decentralized probabilistic reasoning, constraint reasoning, and decision theoretic reasoning are essential tasks of cooperative multiagent systems. Several frameworks for these tasks organize agents into a junction tree (JT). The JT agent organization has a number of computational advantages, including agent privacy during inference computation. During construction of the JT organization, however, a number of existing frameworks utilize construction algorithms that leak the agent privacy, which typically involves proprietary knowledge of the agent developer, including private variables, variables shared between pairs of agents, agent identities and adjacency. One exception is the HTBS algorithm, which constucts a JT organization if one exists without disclosing such private information. A limitation of the HTBS algorithm is that if no JT exists in the given agent environment decomposition, it can only recognize the non-existence. The goal of this thesis research is to extend the HTBS algorithm so that such environmental decomposition can be revised distributively by the agents to enable a JT organization and the computation incurs the minimum loss of agent privacy.