Supplement for high performance compute clusters

数据科学有望为发现提供新的途径，并可以改善 了解、预防和治疗复杂疾病，如癌症、糖尿病、 药物滥用等，这一数字正在显著上升。数据科学的承诺可以是 只有当收集到的数据可以协同共享和分析时，才能充分实现。但 由于不适当的访问以及 越来越多的新型隐私攻击限制了机构共享数据。事实上，在一些 在某些情况下，分析结果本身可能导致重大的隐私损害。成功 的数据共享取决于确保对数据的最大访问， 隐私要求，包括立法规定的要求，以及 组织自己收集数据。虽然现有的解决方案可以解决部分 问题，在真正将这些问题纳入全面工作方面存在重大挑战。 生物医学研究界可用的解决方案，以及 通过现代技术，如深度学习。这项研究的长期目标是开发 能够全面实现数据共享，同时尊重隐私和安全的技术 考虑因素，并确保它们在现有平台中实现， 被研究界广泛接受。为此，本项目的目标是 为风险推理、分布式学习和访问控制开发补充解决方案 可以实现不同形式的数据共享。所研究的问题具有普遍性 并将根据研究的成功和出现的新障碍而发展。拟议 研究计划是重要的，因为缺乏对生物医学数据的访问可能导致 护理分散，导致更高的经济和社会成本，是一个重大的 生物医学研究的障碍。该项目将导致开源，免费提供 软件工具，将被集成到广泛使用的数据收集，队列识别， 分布式分析平台。有几个正在进行的合作，将作为初步的 引导客户提供使用案例、确定需求、评估结果，以及 验证开发的解决方案。

Data science holds the promise of enabling new pathways to discovery and can improve the understanding, prevention and treatment of complex disorders such as cancer, diabetes, substance abuse, etc., which are significantly on the rise. The promise of data science can be fully realized only when collected data can be collaboratively shared and analyzed. However, the widespread increases in healthcare data breaches due to inappropriate access as well as the increasing number of novel privacy attacks restrict institutions from sharing data. Indeed, in some cases, the results of the analysis can themselves lead to significant privacy harm. The success of the data commons depends on ensuring the maximal access to data, subject to all of the patient privacy requirements including those mandated by legislation, and all of the constraints of the organization collecting the data itself. While there are existing solutions that can solve parts of the problem, there are significant challenges in truly incorporating these into comprehensive working solutions that are usable by the biomedical research community, and new challenges brought on by modern techniques such as deep learning. The long-term goal of this research is to develop technologies that can holistically enable data sharing while respecting privacy and security considerations and to ensure that they are implemented in existing platforms that have widespread acceptance in the research community. Towards this, the objective of this project is to develop complementary solutions for risk inference, distributed learning, and access control that can enable different modalities of data sharing. The problems studied are general in nature and will evolve depending on research successes and new impediments that arise. The proposed program of research is significant since lack of access to biomedical data can lead to fragmentation of care, resulting in higher economic and social costs, and is a significant impediment to biomedical research. The project will result in open-source, freely available software tools that will be integrated into widely used data collection, cohort identification, and distributed analytics platforms. There are several ongoing collaborations that will serve as initial pilot customers to provide use cases, identify the requirements, evaluate results, and in general validate the developed solutions.