基于Agent的上下文推理及其在多文化交流平台中的应用

Agent-Based Context Reasoning and Its Application in Intercultural Collaboration Environment

多文化交流平台(intercultural-collaboration-environment,ICE)是亚洲多所大学与研究机构参与的、开发多语种在线交流翻译平台的国际合作项目.针对目前该项目中底层翻译软件AnnoChat翻译质量较差的问题,提出了一种基于Agent的上下文推理机制,以访问领域本体库,辅助选择语义模糊文本的正确意项.通过在ICE中的应用,验证了该机制的有效性.

Intercultural Collaboration Environment（ICE） is an international project on which quite a few Asian universities collaborate. It aims to exploit an online translation platform which enables communication with different Asian languages. Due to the poor comprehensibility of the translation results provided by the underlying translation software AnnoChat which is currently used in ICE project, an Ontology-based model is developed to improve the comprehensibility of translation results of AnnoChat. In the proposed model, this paper foeuses on how agent-based context reasoning is used to pick up the appropriate terms during the translation process. The proposed model was implemented and tested under the context of the ongoing ICE project.