Phase I IUCRC University of Missouri-Kansas City: Center for Big Learning (CBL)

第一阶段 IUCCRC 密苏里大学堪萨斯城分校：大学习中心 (CBL)

• 批准号: 1747751
• 负责人: Zhu Li
• 金额: $ 75万
• 依托单位: University of Missouri-Kansas City
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-02-01 至 2025-01-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1747751&HistoricalAwards=false
• 关键词: Phase; IUCRC; University; Missouri; Kansas

This project establishes the NSF Industry/University Collaborative Research Center (I/UCRC) for Big Learning (CBL) to accelerate innovation and impact of Deep Learning in various embedded applications. The vision is to create intelligence towards intelligence-driven society. Through catalyzing the fusion of diverse expertise from the consortium of faculty members, students, industry partners, and federal agencies, CBL seeks to create state-of-the-art deep learning methodologies and technologies and enable intelligent applications, transforming broad domains, such as business, healthcare, Internet-of-Things, and cybersecurity. This timely initiative creates a unique platform for empowering our next-generation talents with cutting-edge technologies of societal relevance and significance. The University of Missouri at Kansas City (UMKC) site focuses on the deep learning in embedded systems for mobile and IoT applications. It is based on a framework called DeepLite for deep learning model compression and acceleration that can fit cutting edge deep learning capabilities in embedded systems with very limited computing, storage, communication and power capabilities. DeepLite allows embedded deep learning model training and compression for power, storage, computation complexity tradeoffs with learning performances for targeted embedded applications like immersive content capture, depth and action sensing, visual surveillance, next gen image and video compression and communication. CBL is expected to make wide ranging and long lasting impact to machine learning algorithm, system and application research, accelerating deep learning technology innovation and adoption in the real world, enable transformative new capabilities and new applications in all aspect of society, from education, medicine, media, to security and defense. CBL seamlessly integrates innovation, engineering education, technology business incubation, and community engagement. It facilitates closer interactions and cross pollination of ideas between academia and industry, broaden the research horizon for faculties and students, while help shrink the time to impact and time to market of new technology. The center repository will be hosted at http://nsfcbl.org. The data, code, documents will be well organized and maintained on the CBL servers for the duration of the center for more than five years and beyond. The internal code repository will be managed by GitLab. After the software packages are well documented and tested, they will be released and managed by popular public code hosting services, such as GitHub and Bitbucket.

该项目建立了NSF工业/大学合作研究中心（I/UCRC）的大学习（CBL），以加速深度学习在各种嵌入式应用中的创新和影响。我们的愿景是创造智能，走向智能驱动的社会。通过促进教师，学生，行业合作伙伴和联邦机构联盟的各种专业知识的融合，CBL旨在创建最先进的深度学习方法和技术，并实现智能应用，改变广泛的领域，如商业，医疗保健，物联网和网络安全。这一及时的举措创造了一个独特的平台，为我们的下一代人才提供具有社会相关性和重要性的尖端技术。位于堪萨斯城的密苏里州大学（UMKC）网站专注于移动的和物联网应用的嵌入式系统中的深度学习。它基于一个名为DeepLite的框架，用于深度学习模型压缩和加速，可以在计算、存储、通信和电源能力非常有限的嵌入式系统中实现尖端的深度学习功能。DeepLite允许嵌入式深度学习模型训练和压缩，以实现功耗，存储，计算复杂性与目标嵌入式应用程序的学习性能之间的权衡，如沉浸式内容捕获，深度和动作传感，视觉监控，下一代图像和视频压缩以及通信。CBL有望对机器学习算法、系统和应用研究产生广泛而持久的影响，加速深度学习技术在真实的世界中的创新和采用，在社会的各个方面实现变革性的新能力和新应用，从教育、医学、媒体到安全和国防。CBL无缝集成创新，工程教育，技术企业孵化和社区参与。它促进了学术界和工业界之间更密切的互动和思想的交叉授粉，拓宽了教师和学生的研究视野，同时有助于缩短新技术的影响时间和上市时间。中心存储库将托管在http://nsfcbl.org上。数据、代码、文档将在中心的五年以上时间内在CBL服务器上得到良好的组织和维护。内部代码库将由GitLab管理。在软件包经过良好的文档记录和测试后，它们将由流行的公共代码托管服务（如GitHub和Bitbucket）发布和管理。

项目成果

$\lambda$-Domain Perceptual Rate Control for 360-Degree Video Compression

• DOI: 10.1109/jstsp.2019.2963154
• 发表时间: 2020-01
• 期刊: IEEE Journal of Selected Topics in Signal Processing
• 影响因子: 7.5
• 作者: Li Li-Li;Ning Yan;Zhu Li;Shan Liu;Houqiang Li

Network Update Compression for Federated Learning

• DOI: 10.1109/vcip49819.2020.9301815
• 发表时间: 2020-12
• 期刊: 2020 IEEE International Conference on Visual Communications and Image Processing (VCIP)
• 作者: B. Kathariya;Li Li-Li;Zhu Li;Ling-yu Duan;Shan Liu

Advanced CNN Based Motion Compensation Fractional Interpolation

• DOI: 10.1109/icip.2019.8804199
• 发表时间: 2019-09
• 期刊: 2019 IEEE International Conference on Image Processing (ICIP)
• 作者: Hang Zhang;Li Li-Li;Li Song;Xiaokang Yang;Zhu Li

Fast Approximate Score Computation on Large-Scale Distributed Data for Learning Multinomial Bayesian Networks

• DOI: 10.1145/3301304
• 发表时间: 2019-03
• 期刊: ACM Transactions on Knowledge Discovery from Data (TKDD)
• 作者: A. Katib;P. Rao;Kobus Barnard;Charles A. Kamhoua

Mobile Visual Search Compression With Grassmann Manifold Embedding

• DOI: 10.1109/tcsvt.2018.2881177
• 发表时间: 2019-11
• 期刊: IEEE Transactions on Circuits and Systems for Video Technology
• 影响因子: 8.4
• 作者: Zhaobin Zhang;Li Li-Li;Zhu Li;Houqiang Li