The University of Chicago Computer Science Laboratory

该奖项将提供资金，为各种研究开发计算基础设施，主要领域是方程逻辑、基于案例的推理、计算机视觉和图形学、偏微分方程组的数值解以及计算机科学理论。将建立的基础设施将包括一个超大规模集成电路设计设施，供学生和教职员工用于开发和测试电路设计。这一设施将被许多其他研究项目使用。方程式逻辑的研究涉及到方程式逻辑编程的有用实现的发展。将讨论以下问题：编译器优化，基于同余闭包的激进的新实现技术，提供模块化的语言扩展，不具有Church-Rosser性质的不确定系统的实现，保证确定性的更强大的技术，将方程编程与结构编辑器和调试器集成在一起的更灵活的输入输出接口，以及方程程序的并行评估。基于案例推理的研究涉及通过记住先前的结果并使其适应新的情况来分析和实现推理。这种方法将被应用于四个项目：评估学生申请，诊断机械故障，选择送货车辆的路线，以及辅导高中生几何。计算机视觉和图形学的研究涉及到一种新的高效通用的任意方向多面体物体识别算法的研究。将测试算法的实际适用性，并研究算法的普适性。偏微分方程数值解的研究涉及粒子方法、多重网格法、变分不等式和有限元方法中信息的超收敛恢复等问题。这项研究还包括研究用于模拟多孔介质中流体流动的算子分裂和时间分裂方法，以及处理Navier-Stokes方程解的奇异性。研究的理论领域包括复杂性理论和算法、编程语言语义、递归理论，以及递归理论概念和技术在学习的理论基础上的应用。

This award will provide funds to develop computing infrastructure for a wide variety of research, with the primary areas being equational logic, case-based reasoning, computer vision and graphics, numerical solution of partial differential equations, and computer science theory. The infrastructure to be set up will include a VLSI design facility to be used by students and faculty in developing and testing circuit designs. This facility will be used be used by a number of other research projects. The research in equational logic is concerned with the development of useful implementations of an equational logic programming. The following issues will be addressed: compiler optimizations, radical new implementation techniques based on congruence closure, extensions of the language to provide modularity, implementation of indeterminate systems without the Church-Rosser property, more powerful techniques for guaranteeing determinacy, more flexible input-output interfaces integrating equational programming with structure editors and debuggers, and parallel evaluation of equational programs. The research on case-based reasoning is concerned with the analysis and implementation of reasoning by remembering previous results and adapting them to new situations. This approach will be applied to four projects: evaluating student applications, diagnosing mechanical failures, routing delivery vehicles, and tutoring high school students in geometry. The research on computer vision and graphics involves the study of a new efficient general purpose algorithm for recognizing polyhedral objects in arbitrary orientations. Practical applicability of the algorithm will be tested and generalizations of the algorithm will be investigated. The research on numerical solutions of partial differential equations involves several projects dealing with particle methods, multigrid methods, variational inequalities, and superconvergent recovery of information in finite element methods. The research also involves studying operator-splitting and time-splitting methods for the simulation of fluid flow in porous media, and dealing with singularities in solutions of the Navier-Stokes equations. The theoretical areas of research include complexity theory and algorithms, programming language semantics, recursion theory, and the application of recursion theoretic concepts and techniques to the theoretical foundations of learning.