Workshop on Advanced Cyber-Enabled Discovery & Innovation (CDI) Through Symbolic and Numeric Computation

符号计算学科为基于网络的发现和创新提供了先进的软件。符号方法和程序被应用于科学计算、物理、数学、系统间通信协议和标准、范畴语言设计、泛型编程和编译等领域的问题解决和模型发现。由于计算的复杂性，并行和分布式环境开始使用，符号和数值方法结合成混合方法，并进行语义验证和输出验证(例如，通过拉斯维加斯随机化和数值误差分析)。我们计划于2007年10月底或11月初在华盛顿特区聚集24-30名来自学术界和工业界的国际公认的符号计算及其CDI应用学科(如数值计算、物理、数学、计算语言设计和编译、并行和分布式计算、程序接口设计)的专家小组，为期一天半的研讨会。我们研讨会的目标是讨论和编写一份报告，说明通过符号计算对CDI可能的研究方向及其协同作用，包括在符号计算和应用学科中的研究方向。符号计算的新用途可以是符号/数值混合方法的稀疏模型构建、物理和数学问题的稀疏精确矩阵计算、多项式系统解和非线性优化的多级并行软件设计、高级分类和可验证的程序设计、算法综合等。

The discipline of symbolic computation provides advanced software for cyber-enabled discovery and innovation. Symbolic methodologies and programs are applied to problem solving and model discovery in--scientific computing,--physics,--mathematics,--inter-system communication protocols and standards,--categorical language design, generic programming and compilation,and more. Due to the complexity of the arising computations, parallel and distributes environments are coming into use, symbolic and numeric approaches are combined into hybrid methods, and semantic verification and output certification (e.g., through Las Vegas randomization and numerical error analysis) are performed.We propose to gather in late October or early November 2007 in Washington DC a group of 24--30 invited internationally recognized experts from academia and industry in both symbolic computation and its CDI applications disciplines (e.g., numerical computing, physics, mathematics, computing language design and compilation, parallel and distributed computation, program interface design) for a one and one half day workshop. The objectives of our workshop are to discuss and produce a report on possible research directions and their synergies for CDI through symbolic computation, both in the symbolic computation and the applications disciplines. Novel uses of symbolic computation can be sparse model construction by hybrid symbolic/numeric methods, sparse exact matrix computations for problems from physics and mathematics, multi-level parallel software design for polynomial system solution and non-linear optimization, high-level categorical and verifiable program design, algorithm synthesis, and more.