EAGER: DBfN: Data Bridge for Neuroscience: A novel way of discovery for Neuroscience Data

在数以千计的科学家正在创建数百万个描述神经科学现象日益多样化的数据集的时候，一名科学家在没有帮助的情况下找到与特定研究领域相关的所有数据的可能性每天都在下降。与此同时，存储数据数量和多样性的快速增长意味着这些数据集为重要的新合作研究和发现提供支持的潜力也相应增加。应对这些挑战需要专门创建的工具来帮助科学家寻找相关的数据集和合作者。由个别科学家和团队收集和生成的大量相对较小的数据集，形成了一类独特的大数据，称为“科学的长尾”数据，利用它们的隐藏力量对于推进科学至关重要。该项目将参与一系列计划活动，以应用于神经科学的新型数据发现系统，基于一个名为DataBridge的平台，该平台由研究人员及其合作者在NSF大数据计划的资助下开发。DataBridge应用“签名”和“相似性”算法在语义上将大量不同的数据集连接到一个“社会计量”网络中。之前的工作主要集中在社会科学的数据上。这项工作将研究将这些技术扩展到神经科学领域。如果这些技术能够被证明适用于神经科学数据，该项目不仅将在神经科学研究界产生广泛影响，而且可能还会在其他科学界产生广泛影响。神经科学正处于一个转折点，越来越多的数据正在通过公共存储库进行汇总和共享。在大数据时代，神经科学界面临的主要挑战是在这些存储库中发现相关数据集的难度。有效发现和识别相关数据的复杂性构成了科学数据长尾的最后一英里问题。解决这一问题可以通过重复使用和重新调整用途来增加数据的价值，并可以通过增加对不同主题领域的数据的访问，极大地使国家科学基金会大脑倡议受益。该项目将基于一个名为DataBridge的平台启动对神经科学的新型数据发现系统的应用研究，该平台是该项目团队在NSF大数据计划的资助下开发的。DataBridge应用“签名”和“相似性”算法在语义上将大量不同的数据集连接到一个“社会计量”网络中。神经科学数据桥平台将尝试利用神经科学研究人员开发的复杂分析算法，以便从大量和不同的所谓长尾神经科学数据中提取关键签名并找到数据关联。通过提供一个场所，通过模式分析、特征提取和其他相关标准来定义复杂的搜索标准，DataBridge为科学数据提供了一个高度可定制的搜索引擎。该项目将对DataBridge在神经科学数据上的适用性进行初步的可行性研究，目标有两个：1.为神经科学实施一个试验性的DataBridge系统，并演示从语义上连接一小批神经科学数据集的概念验证；2.举办一次研讨会，发展神经科学界的用户联盟，以便建立一个可持续的以DataBridge为基础的神经科学基础设施。这项基于社区的活动将利用NSF大数据中心和Spokes计划创建的社区基础设施。

At a time when thousands of scientists are creating millions of datasets describing an increasingly diverse mix of neuroscience phenomena, the chances of an individual unaided scientist finding all of the data relevant to a particular line of investigation are shrinking every day. At the same time, this rapid increase in the amount and diversity of stored data implies a corresponding increase in the potential of these datasets to empower important new collaborative research and discovery. Meeting these challenges requires tools specifically created to assist scientists in their search for relevant datasets and collaborators. Massive number of relatively small datasets gathered and generated by individual scientists and groups, form a distinct class of Big Data called the "long tail of science" data and harnessing their hidden power is crucial for advancing science. This project will engage in a set of planning activities for the application of a novel data discovery system for Neuroscience, based upon a platform called, DataBridge, which has been developed by the investigators and their collaborators under a grant from the NSF Big Data program. DataBridge applies "signature" and "similarity" algorithms to semantically bridge large numbers of diverse datasets into a "sociometric" network. Prior work has focused on data from the Social Sciences. This work will study the extension of those techniques to the Neuroscience domain. If the techniques can be demonstrated to work with neuroscience data, the project would have broad impact not only in the neuroscience research community but, potentially, in other science communities as well.Neuroscience is at an inflection point where more and more data are being aggregated and shared through common repositories. The main challenge facing the neuroscience community in the Big Data era is the difficulty of discovering relevant datasets across these repositories. The complication of effective discovery and identification of relevant data forms the last mile problem for long tail of science data. Solving this problem can increase the value of the data through reuse and repurposing and can immensely benefit the NSF Brain Initiative by providing increased access to data in its various thematic areas. This project will initiate studies on the application of a novel data discovery system for Neuroscience based upon a platform called DataBridge, which the project team has developed under a grant from the NSF Big Data program. DataBridge applies "signature" and "similarity" algorithms to semantically bridge large numbers of diverse datasets into a "sociometric" network. The DataBridge for Neuroscience platform will attempt to harness complex analytics algorithms developed by neuroscience researchers in order to extract key signatures and find data associations from large volumes, and diverse collections, of so-called "long-tail" neuroscience data. By providing a venue for defining complicated search criteria through pattern analysis, feature extraction and other relevance criteria, DataBridge provides a highly customizable search engine for scientific data. This project will conduct a preliminary, feasibility study on the applicability of DataBridge on Neuroscience data with two goals: 1. Implement a pilot DataBridge system for Neuroscience and demonstrate a proof of concept for semantically bridging a small collection of neuroscience datasets, and 2. Conduct a workshop to develop a coalition of users from the neuroscience community in order to build a sustainable DataBridge-based infrastructure for neuroscience. This community-based activity will leverage the community infrastructure created by the NSF Big Data Hubs and Spokes program.