CAREER: Storing, Querying and Re-Using Provenance of Computational Tasks

基于工作流的系统已经成为在科学界广泛使用的数据探索的一种替代方法。工作流可以捕获各种细节级别的计算任务，并系统地记录再现性、结果发布和共享所需的来源（历史）信息。虽然使用科学工作流系统的好处是众所周知的，但工作流难以创建和维护的事实已经成为科学领域广泛采用该技术的主要障碍。这个项目的目标是产生新的算法和技术，用于探索和重用嵌入在工作流规范和它们操作的数据来源中的有用知识。该项目解决了现有工作流系统中的关键限制。首先，它开发了一组可用的工具，使临时用户（不一定具有编程专业知识）能够执行探索性任务并通过工作流解决问题。这些包括直观的用户界面，用于操作工作流集合和通过示例查询工作流。其次，它构建了一个可扩展的来源管理基础设施，以支持这些操作的有效执行。该项目的研究成果推动了当前技术的发展，并在存储、查询和重用计算任务的来源方面建立了基础知识。这个项目有可能影响到各种各样的应用程序，其中工作流的创建和维护目前是一个主要的瓶颈。这包括大型计算科学项目和门户。此外，它使工作流和工作流技术更容易被普通用户访问。通过我们的跨学科合作，该项目将对帮助改进科学发现过程产生直接影响。研究生和本科生参与该项目将提供指导机会。PI致力于招收少数民族学生。该项目的结果将作为研究论文和免费工具在项目网站：http://www.cs.utah.edu/~juliana/projects/NSF-IIS-0746500上发布

Workflow-based systems have emerged as an alternative to ad-hocapproaches to data exploration that are widely used in the scientificcommunity. Workflows can capture computational tasks at various levelsof detail and systematically record the provenance (history) informationnecessary for reproducibility, result publication andsharing. Although the benefits of using scientific workflow systemsare well known, the fact that workflows are hard to create andmaintain has been a major barrier to wider adoption of the technologyin the scientific domain.The goal of this project is to produce new algorithms and techniquesfor exploring and re-using useful knowledge embedded in workflowspecifications and in the provenance of the data they manipulate.This project addresses key limitations in existing workflowsystems. First, it develops a set of usable tools that enable casualusers (who do not necessarily have programming expertise) to performexploratory tasks and solve problems through workflows. These includeintuitive user interfaces to manipulate collections of workflow and toquery workflows by example. Second, it builds a scalable provenancemanagement infrastructure to support the efficient execution of theseoperations.The research results of this project advance the state of the art andbuild fundamental knowledge in storing, querying, and re-usingprovenance of computational tasks. This project has the potential toimpact a variety of applications where the creation and maintenance ofworkflows is currently a major bottleneck. This includes largecomputational science projects and portals. Furthermore, it makesworkflows and workflow technology more accessible to casual users.Through our interdisciplinary collaborations, this project will haveimmediate impact in helping improve the scientific discovery process.The involvement of graduate and undergraduate students in the projectwill provide mentoring opportunities. The PI is committed torecruiting minority students. The results of this project will bedisseminated as research papers and as freely available tools at theproject website:http://www.cs.utah.edu/~juliana/projects/NSF-IIS-0746500