IIS: Knowledge Representation and Reasoning Mechanisms for Explicitly Self-Aware Communicative Agents

IIS：显式自我意识交流代理的知识表示和推理机制

• 批准号: 0535105
• 负责人: Lenhart Schubert
• 金额: $ 29.94万
• 依托单位: University of Rochester
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-07-01 至 2009-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0535105&HistoricalAwards=false
• 关键词: IIS; Knowledge; Representation; Reasoning; Mechanisms

This project is developing theoretical and practicalfoundations for communicative agents that overtly exhibitself-awareness. Such agents are aware of what they knowand do not know, what they can and cannot do, and whatintentions and assumptions underlie their inferencesand actions, and they show "presence of mind" in dialogueinteractions, recalling past experiences and reasoningabout the ongoing interactions. The theoretical ideasunder development for this purpose center around theoriesof autocognitive knowledge and inference (knowing whatone knows and does not know, and how one acquiresknowledge), a new approach to integrating uncertaininference with logic (quasi-probability theory), anda form of "syntactically aware" metainference thatfacilitates reasoning about a system's own knowledgeand special capabilities. The project uses the EPILOGinference engine as a basis for experimentation, whilebroadening that engine's capabilities. The knowledgerepresentation (episodic logic) used by EPILOG nearlymatches the expressive capabilities of ordinary languageand is thus well-suited to the inferential and dialoguegoals of this project. The results will includedemonstrable theories of self-awareness in artificialagents, new methods of general autocognitive, uncertain,and syntactically aware inference, and an improvedEPILOG engine. The results will be disseminated in majorconferences and journals for AI, agent and dialogueresearch, and the improved EPILOG engine will begenerally available. In the longer run, the work willhelp pave the way for a new generation of more human-likeinteractive AI agents, in such areas as software support,advice, tutoring, games, etc.

这个项目正在为公开展示自我意识的交际体发展理论和实践基础。这些智能体意识到他们知道什么和不知道什么，他们能做什么和不能做什么，以及他们的推理和行动背后的意图和假设，他们在对话互动中表现出“头脑的存在”，回忆过去的经历并对正在进行的互动进行推理。为此目的正在发展的理论思想围绕着自动认知知识和推理理论（知道自己知道什么和不知道什么，以及如何获得知识），一种将不确定性推理与逻辑（准概率论）相结合的新方法，以及一种促进对系统自身知识和特殊能力进行推理的“句法感知”元影响形式。该项目使用EPILOGinference引擎作为实验的基础，同时扩展了该引擎的功能。EPILOG使用的知识表示（情景逻辑）几乎与普通语言的表达能力相匹配，因此非常适合该项目的推理和对话目标。结果将包括人工智能体中自我意识的可论证理论，一般自动认知，不确定和句法意识推理的新方法，以及改进的depilog引擎。结果将在人工智能、代理和对话研究的主要会议和期刊上发布，并且改进的EPILOG引擎将普遍可用。从长远来看，这项工作将有助于在软件支持、咨询、辅导、游戏等领域为新一代更像人类的交互式人工智能代理铺平道路。