Presidential Young Investigator Awards

这个研究项目是关于变化推理的问题。在实践方面，支持工程信息系统（EIS）变更的方法，即关于工程工件（如VLSI电路或软件系统）的信息存储库将被研究。现在人们普遍认识到，信息管理对大规模设计工作的成功至关重要。该领域的近期项目包括构建基于主存储器的数据管理器，用于VLSI电路的高级合成；调查EIS更新的性能问题，特别是当同一设计的不同方面由不同的工程师并行更新时，导致数据一致性和任务协调问题，这在大型工程项目中至关重要。从更理论化的角度来看，人们对逻辑理论变化的形式推理方法很感兴趣。例如，当一个关于世界的新信息被发现时，人们如何决定保留哪些旧的知识和数据，以及根据新信息丢弃哪些知识和数据？这项工作涉及数据库（更新不完全信息的数据库）、哲学（信念修正、科学理论修正）和人工智能（规划、诊断、行为效果推理等）中的问题。理论修正的研究将强调AI的问题，如对房间中物体位置的推理，作为修正技术的应用。最近关于理论修正与限制和行为推理的关系的研究已被证明是特别富有成效的，这些研究将进一步进行。

This research project is concerned problems associated with reasoning about change. On the practical side, means of supporting changes in an Engineering Information System (EIS), that is a repository of information about an engineering artifact such as a VLSI circuit or software system will be investigated. It is now widely recognized that information management is crucial to the success of large scale design efforts. Near term projects in this area include construction of a main memory based data manager for high level synthesis of VLSI circuits; investigation of performance issues for updates in EIS, particularly as different aspects of the same design are updated in parallel by different engineers, leading to problems of data consistency and task coordination that are critical in large engineering projects. From a more theoretical perspective, there is interest in formal means of reasoning about changes in logical theories. For example, when a new piece of information about the world is discovered, how can one decide what old pieces of knowledge and data to keep, and what to discard in the light of the new information? This work is relevent to problems in databases (updating databases with incomplete information), philosophy (belief revision, revision of scientific theories), and artificial intelligence (planning, diagnosis, reasoning about the effects of actions, etc.). The studies of theory revision will emphasize problems from AI, such as reasoning about the location of objects in a room, as applications for revison techniques. Recent studies of the relationship of theory revision to circumscription and reasoning about action have proven particularly fruitful, and these will be pursued further.