University of Missouri-Kansas City Planning Grant: I/UCRC for Big Learning

拟建的NSF I/UCRC大学习中心(CBL)的使命是探索新兴大规模深度学习(DL)的研究前沿，以实现有效和高效的计算智能，设计新的学习算法和系统机制，用于大数据和大系统时代的智能研究和应用。通过由多个学术站点(与佛罗里达、CMU和俄勒冈合作)和大量行业合作伙伴组成的大型学习联盟，该中心寻求催化学术界、政府、行业利益相关者的智慧融合，算法、系统和教育方面的快速创新，以及将技术转移到具有现实世界相关性和意义的尖端产品和服务中。拟建中心的更广泛影响：随着自然系统、工程系统和人类/生命活动产生的数据爆炸性增长，我们需要智能软件和硬件来促进我们的决策，并在规模上自动提取洞察力。在我们的社会迈向充满智能的机会世界之际，拟议的I/UCRC大学习中心是一项及时的倡议。大学习联盟预计将成为深度学习研究和应用的磁石，并吸引领先的研究人员、热情的企业家、IT和行业巨头共同努力，实现CBL充满希望的使命和愿景。特别是，CBL具有以下更广泛的影响。(1)在开创性研究和应用方面对深度学习社区做出重大贡献和影响，以应对广泛的现实世界挑战。(2)对推广工业界的产品和服务，特别是我们会员的产品和服务作出重大贡献和影响。(3)为我们迫切需要的下一代人才的教育做出重大贡献和影响，这些人才来自真实世界的背景和来自学术界和产业界的世界级导师。(4)我们的会议、论坛、会议和计划中的培训课程将极大地促进和拓宽DL的研究和实现。随着在多种形式的挑战(如图像、视频、语音、文本和生命、健康和科学数据)的信号压缩、分类和识别方面取得重大突破，计算智能的复兴正在迫在眉睫。CBL的使命是通过团结一致的努力以及深度整合和融合我们大量教职员工、学生和行业合作伙伴的广泛专业知识，引领这一新兴趋势。CBL的愿景是为智能驱动的社会创造智能使能器。CBL在以下关键研究主题方面具有开创性的智力价值。(1)算法新颖。本文重点介绍了深度学习算法和体系结构，如深度结构、复杂深度神经网络、脑启发组件、深度学习的优化和加速、神经机器以及对传统机器学习算法的适应。(2)制度创新。我们为嵌入式设备、移动设备、台式机、集群和云提出了新颖的资源管理策略、不同的架构和软件工具包。(3)在商业、健康、成像和智能物方面的新应用，包括在新的图像/视频建模和压缩中的深度残差网络，在熵编码中的大规模上下文模型的RNN，大规模视觉对象的重新识别，以及利用成像的靶向药物输送。在规划阶段，我们将建立扎实的中心战略规划、营销规划，以及由五个学术站点和几十个行业成员组成的大学习联合体。

The mission of the proposed NSF I/UCRC Center for Big Learning (CBL) is to explore research frontiers in emerging large-scale deep learning (DL) to realize effective and efficient computational intelligence, design novel learning algorithms and system mechanisms for intelligence research and applications in the era of big data and big systems. Through the big learning consortium of multiple academic sites (in collaboration with Florida, CMU, and Oregon) and a large number of industry partners, the center seeks to catalyze the fusion of wisdom from academia, government, industry stakeholders, the rapid innovation in algorithms, systems, and education, and technology transfer into cutting-edge products and services with real-world relevance and significance. Broader Impacts of the proposed center: with the explosive growth of data generated from natural systems, engineered systems, and human/life activities, we need intelligent software and hardware to facilitate our decision making with distilled insights automatically at scale. The proposed I/UCRC Center for Big Learning is a timely initiative as our society moves towards intelligence-enabled world of opportunities. The Big Learning consortium is expected to become the magnet of deep learning research and applications and attract leading researchers, enthusiastic entrepreneurs, IT and industry giants working together on accomplishing the promising missions and visions of CBL. In particular, CBL has the following broader impacts. (1) Making significant contributions and impacts to the deep learning community on pioneering research and applications to address a broad spectrum of real-world challenges. (2) Making significant contributions and impacts to promote products and services of industry in general and our members in particular. (3) Making significant contributions and impacts to the urgently needed education of our next-generation talents with real-world settings and world-class mentors from both academia and industry. (4) Our meetings, forums, conferences, and planned training sessions will greatly promote and broaden the research and materialization of DL.With dramatic breakthroughs in signal compression, classification and identification in multiple modalities of challenges (e.g., image, video, speech, text, and life, health & science data), the renaissance of computational intelligence is looming. The mission of the CBL is to pioneer in this emerging trend through united and coordinated efforts and deep integration and fusion of broad expertise from our large number of faculty members, students, and industry partners. The vision of CBL is to create intelligence enablers towards intelligence-driven society. CBL possesses the pioneering intellectual merit in the following key research themes. (1) Novel algorithms. This theme focuses on novel DL algorithms and architectures, such as deep architectures, complex deep neural networks, brain-inspired components, optimization and acceleration of the deep learning, neural machines, and adaptation of conventional machine learning algorithms. (2) Novel systems. We propose novel resource management strategies, heterogeneous architectures, and software tool kits for embedded devices, mobiles, desktops, clusters, and clouds. (3) Novel applications in business, health, imaging, and smart things, including deep residual networks in new image/video modeling and compression, RNN for large scale context models in entropy coding, large scale visual object re-identification, and targeted drug delivery with imaging. During the planning phase, we will establish a solid center strategic plan, marketing plan, and the consortium of big learning that consists of five academic sites and several dozens of industrial members.