Similarity Search for Scientific Workflows
Similarity Search for Scientific Workflows
复制标题
DOI：
10.14778/2732977.2732988
复制
发表时间：
2014-08
期刊：
影响因子：
--
通讯作者：
J. Starlinger;Bryan Brancotte;Sarah Cohen Boulakia;U. Leser
中科院分区：
文献类型：
--
作者：
J. Starlinger;Bryan Brancotte;Sarah Cohen Boulakia;U. Leser
随着科学工作流的日益流行，公共存储库作为共享、查找和重复使用此类工作流的一种手段正变得越来越重要。随着这些存储库的大小增长，比较存储在其中的科学工作流的方法变得必要，例如，允许重复检测或相似性搜索。科学工作流是复杂的对象，它们的比较需要许多不同的步骤，从比较原子元素到比较整个工作流。各种研究已经实现了科学工作流比较的方法，并得出了经常相互矛盾的结论，其中哪些算法工作得最好。比较这些结果很麻烦，因为最初的研究混合了不同步骤的不同方法，并使用了不同的评估数据和衡量标准。我们对该领域的贡献是：(I)通过将每个先前的方法分解成明确定义的和可比较的子任务集，(Ii)通过孤立地比较在科学工作流比较的每个步骤中采取的不同方法，报告一些意想不到的发现，(Iii)通过调查如何将这些方法最好地结合到聚合度量中，以及(Iv)通过提供由15个工作流专家在近1500个工作流和我们评估的所有方法的重新实现的语料库上贡献的2000多个相似性评级的黄金标准来做出贡献。
With the increasing popularity of scientific workflows, public repositories are gaining importance as a means to share, find, and reuse such workflows. As the sizes of these repositories grow, methods to compare the scientific workflows stored in them become a necessity, for instance, to allow duplicate detection or similarity search. Scientific workflows are complex objects, and their comparison entails a number of distinct steps from comparing atomic elements to comparison of the workflows as a whole. Various studies have implemented methods for scientific workflow comparison and came up with often contradicting conclusions upon which algorithms work best. Comparing these results is cumbersome, as the original studies mixed different approaches for different steps and used different evaluation data and metrics. We contribute to the field (i) by disecting each previous approach into an explicitly defined and comparable set of subtasks, (ii) by comparing in isolation different approaches taken at each step of scientific workflow comparison, reporting on an number of unexpected findings, (iii) by investigating how these can best be combined into aggregated measures, and (iv) by making available a gold standard of over 2000 similarity ratings contributed by 15 workflow experts on a corpus of almost 1500 workflows and re-implementations of all methods we evaluated.