Advanced Algorithms for Plan Recognition

先进的计划识别算法

• 批准号: RGPIN-2017-06350
• 负责人: Kabanza, Froduald
• 金额: $ 2.48万
• 依托单位: Université de Sherbrooke
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2022
• 资助国家: 加拿大
• 起止时间: 2022-01-01 至 2023-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=757301
• 关键词: Advanced; Algorithms; Plan; Recognition

Artificial intelligence (AI) presents the potential to profoundly transform our society with various applications that will positively affect the economy and our everyday life. Intelligent automated agents will soon be driving applications in many including, transportation, social media, game AI, educational tools, intelligence analysis, defense and security. The ability of these agents to effectively support and naturally interact with people will largely depend on their proficiency to recognize the goals and plans of other agents. The problem of recognizing the goals or plans of other agents is known as the plan recognition problem or the intent recognition problem. While interesting advances have been made over the years, today plan recognition algorithms are too ineffective for many applications. The objective of this research program is to explore, improve, and develop plan recognition algorithms that can cope with complex applications. The research avenues articulate around automated planning, algorithmic game-theory and machine learning.

人工智能（AI）具有深刻改变我们社会的潜力，其各种应用将对经济和我们的日常生活产生积极影响。智能自动化代理将很快推动许多应用，包括交通，社交媒体，游戏AI，教育工具，情报分析，国防和安全。这些代理人有效支持和自然地与人互动的能力在很大程度上取决于他们识别其他代理人的目标和计划的熟练程度。识别其他主体的目标或计划的问题被称为计划识别问题或意图识别问题。虽然这些年来已经取得了有趣的进展，但今天的计划识别算法对于许多应用来说太无效了。该研究计划的目标是探索，改进和开发可以科普复杂应用的计划识别算法。研究途径围绕自动规划，算法博弈论和机器学习。