Multiagent Distributed Ontology Learning

Workshop on Ontologies in Agent Systems, Bologna

计算机科学;社会科学（总论）

Leen-Kiat Soh

Others  :  http://CEUR-WS.org/Vol-66/oas02-5.pdf
PID  :  18689

来源: CEUR

【 摘 要 】

In this paper, we describe a framework for distributed ontologylearning embedded in a multiagent environment.The objective ofthis framework is to improve communication and understandingamong the agents while preserving agent autonomy.Each agentmaintains a dictionary for its own experience and a translationtable.The dictionary allows the agent to compare and discoverrelationships between a pair of words or concepts, while thetranslation table enables the agent to learn and record (a selectedportion of) the vocabulary of its neighbors that is useful for thecollaboration among the agents.The motivation for this distrib- uted ontology learning is that each agent has its own experienceand thus learns its own ontology depending on what it has beenexposed to.As a result, different agents may use different wordsto represent the same experience.When two agents communi- cate, agent A may not understand what agent B and that hinderscollaboration.However, equipped with the distributed ontologylearning capabilities, agents are able to evolve independently theirown ontological knowledge while maintaining translation tables through learning to help sustain the collaborative effort.