MRI: Acquisition of Artificial Intelligence & Deep Learning (AIDL) Training and Research Laboratory

MRI：人工智能的获取

• 批准号: 1828181
• 负责人: Xingquan Zhu
• 金额: $ 65.29万
• 依托单位: Florida Atlantic University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-10-01 至 2021-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1828181&HistoricalAwards=false
• 关键词: MRI; Acquisition; Artificial; Intelligence; &amp

Researchers in health, biomedical science, and various engineering fields often do not receive sufficient training in using the most powerful approach to machine learning known to date, an approach called "deep learning", for analying their data. Deep learning is based on simulated artificial neural networks, and for large real-world problems requires access to specialized computer hardward and software. Such hardware and software platforms, however, are rarely part of the information technology resources available for researchers outside of the field of computer science. This project will overcome this barrier by procurement and development of a deep learning platform at Florida Atlantic University for such research. The project provides a training hub for industry and university to work closely on advanced artificial intelligence applications, and in turn might benefit economic growth.This infrastructure project supports creation of a deep learning platform for health, biomedicine, ocean research, and related domains at Florida Atlantic University. The platform will be shared across campus to service multiple domains. The project brings about a centralized cross campus interdisciplinary platform and augmented deep learning and related artificial intelligence tools for interdisciplinary research. The former, jointly managed by the College of Engineering and Computer Science, and the FAU office of Information Technology, enables building upon the experience and frameworks of others that eventually results in shared infrastructure savings. The latter is likely to contribute in building/augmenting AI and DL tool kits for interdisciplinary research, including augmentation of existing common machine learning and deep learning algorithms for domain experts to carry out analysis on their data without requiring intensive programming skills. Augmented AI and DL tools should be particularly useful to ocean engineers and health sciences.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

健康、生物医学和各种工程领域的研究人员通常没有接受足够的培训，无法使用迄今为止已知的最强大的机器学习方法（一种称为“深度学习”的方法）来分析他们的数据。 深度学习基于模拟人工神经网络，对于大型现实问题需要使用专门的计算机硬件和软件。 然而，此类硬件和软件平台很少成为计算机科学领域之外的研究人员可用的信息技术资源的一部分。 该项目将通过在佛罗里达大西洋大学采购和开发用于此类研究的深度学习平台来克服这一障碍。 该项目为工业界和大学提供了一个培训中心，以便在先进人工智能应用方面密切合作，从而可能有利于经济增长。该基础设施项目支持在佛罗里达州创建一个针对健康、生物医学、海洋研究和相关领域的深度学习平台大西洋大学。 该平台将在校园内共享，为多个领域提供服务。该项目带来了一个集中的跨校园跨学科平台以及用于跨学科研究的增强深度学习和相关人工智能工具。 前者由工程与计算机科学学院和 FAU 信息技术办公室共同管理，能够借鉴其他人的经验和框架，最终实现共享基础设施的节省。后者可能有助于构建/增强跨学科研究的人工智能和深度学习工具包，包括增强现有的通用机器学习和深度学习算法，以便领域专家无需密集的编程技能即可对其数据进行分析。增强型人工智能和深度学习工具应该对海洋工程师和健康科学特别有用。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Learning Graph Neural Networks with Positive and Unlabeled Nodes

学习具有正节点和未标记节点的图神经网络

• DOI: 10.1145/3450316
• 发表时间: 2021-03-08
• 期刊: ACM Transactions on Knowledge Discovery from Data (TKDD)
• 作者: Man Wu;Shirui Pan;Lan Du;Xingquan Zhu
• 通讯作者: Xingquan Zhu

OpenWGL: Open-World Graph Learning

OpenWGL：开放世界图学习

• 发表时间: 2020-01
• 期刊: Italy
• 作者: Wu, Man;Pan, Shirui;Zhu, Xingquan
• 通讯作者: Zhu, Xingquan

Predictive Modeling of Hospital Readmission: Challenges and Solutions

再入院的预测模型：挑战与解决方案

• DOI: 10.1109/tcbb.2021.3089682
• 发表时间: 2021-01
• 期刊: IEEE/ACM Transactions on Computational Biology and Bioinformatics
• 作者: Wang, Shuwen;Zhu, Xingquan
• 通讯作者: Zhu, Xingquan

GAEN: Graph Attention Evolving Networks

GAEN：图注意力进化网络

• DOI: 10.24963/ijcai.2021/213
• 发表时间: 2021-08
• 期刊: Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI
• 作者: Shi, Min;Huang, Yu;Zhu, Xingquan;Tang, Yufei;Zhuang, Yuan;Liu, Jianxun
• 通讯作者: Liu, Jianxun

Dropout vs. batch normalization: an empirical study of their impact to deep learning

Dropout 与批量归一化：对深度学习影响的实证研究

• DOI: 10.1007/s11042-019-08453-9
• 发表时间: 2020-05
• 期刊: Multimedia Tools and Applications
• 影响因子: 3.6
• 作者: Garbin, Christian;Zhu, Xingquan;Marques, Oge
• 通讯作者: Marques, Oge