学术视点

学术视点

日期:2025-12-29

|  来源:【字号:

题目:A Comprehensive Survey of Federated Open-World Learning(联邦开放世界学习综合综述)

作者:Z. Cai, J. Pang, Y. Li, Y. Huang and Z. Xie

来源:IEEE Transactions on Network Science and Engineering(网络科学与工程汇刊), vol. 13, pp. 208-224

摘要:The rapid development of large-scale, diverse data and machine learning (ML) technologies has facilitated the rise of intelligent applications across various sectors. However, concerns over privacy and security risks associated with data collection and ML model training have prompted the emergence of Federated Learning (FL), a distributed machine learning paradigm that ensures privacy-preserving capabilities through collaborative model training. Initially applied to Google's Gboard, FL has since found widespread adoption in domains such as intelligent transportation, recommendation systems, and healthcare. Despite its success, existing FL models primarily focus on optimizing performance while safeguarding privacy, often overlooking the collaborative group's ability to adapt to environmental changes. Drawing parallels to human societies, which effectively adapt to both individual and collective changes, we propose that FL's adaptation to dynamic environments can enhance its ability to support real-world applications. This paper introduces the concept of Federated Open-World Learning (FOWL), a framework that enables FL to not only respond to changes but learn patterns from them, addressing challenges such as participant variability, multi-task learning, and catastrophic forgetting. We provide an evolving view of FL techniques, discuss the shift towards open-world conditions, and compare existing methods for implementing FOWL. Finally, we highlight a number of key challenges and potential future research directions for advancing federated learning in dynamic, open-world environments.

编者译:随着大规模多样化数据和机器学习技术的快速发展,智能应用在各行各业迅速兴起。然而,数据收集与模型训练所带来的隐私安全风险引发广泛关注,推动了联邦学习这一分布式机器学习范式的兴起。联邦学习通过协同训练实现隐私保护,最初应用于谷歌Gboard输入法,现已广泛应用于智能交通、推荐系统、医疗健康等领域。然而,现有联邦学习模型主要聚焦于隐私保护前提下的性能优化,却往往忽视了其对动态环境的协同适应能力。受人类社会既能适应个体变化也能应对集体变迁的启发,本文认为联邦学习对动态环境的适应能力将进一步提升其在实际场景中的应用价值。本文提出联邦开放世界学习的概念,使联邦学习不仅能够响应环境变化,更能从中学习规律,以应对参与者异质性、多任务学习与灾难性遗忘等挑战。本文系统梳理了联邦学习技术的演进脉络,探讨了开放世界条件的转变,并比较了现有的联邦开放世界学习实现方法。最后,本文指出了该领域面临的关键挑战,并展望了未来在动态开放环境下推动联邦学习发展的研究方向。

题目:面向韧性社会的信息韧性理论内涵与实践路径

作者:邓胜利、袁梦

来源:图书情报知识, 2025, 42(5): 44-53,65.

摘要:本文围绕韧性社会发展的核心需求,立足应对危机风险,深度剖析韧性社会与信息韧性二者之间的内在联系,对信息韧性概念进行界定并梳理其构成要素。信息韧性涵盖信息制度韧性、信息技术韧性和信息治理韧性三重内容;从理论逻辑角度揭示信息韧性如何驱动韧性社会发展,在风险防控的承受、抵御、转化三阶段形成递进式作用链条。就韧性社会现面临的风险状况提出三大路径:信息政策动态调整、智能技术赋能、虚假信息纠查与纠偏,以期解决我国韧性社会构建的现实难题与痛点,全面提升我国社会的信息韧性水平。

题目:Terahertz Sensing, Communication, and Networking: A Survey(太赫兹的感知、通信与网络综述)

作者:H. Zhang et al.

来源:IEEE Transactions on Network Science and Engineering(网络科学与工程汇刊), vol. 13, pp. 501-521.

摘要:Terahertz has emerged as a pivotal technology that will enable next-generation wireless systems due to its unique properties. Over the past few years, numerous studies related to terahertz have been conducted, prompting us to undertake a comprehensive survey to summarize the sensing, communication, and networking technologies based on terahertz. In this survey, we first introduce the characteristics of terahertz and its associated hardware devices. Subsequently, we provide a comprehensive review of existing terahertz-based sensing, communication, and networking technologies. We analyze the key techniques, system design, and application scenarios. Finally, we discuss the challenges and future directions of terahertz technology.

编者译:太赫兹技术因其独特性质,已成为推动下一代无线系统发展的关键技术。本文首先介绍了太赫兹的技术特性及相关硬件设备,继而全面梳理了现有太赫兹感知、通信与网络技术的研究进展,分析了其关键技术、系统设计与应用场景,最后探讨了太赫兹技术面临的挑战与未来发展方向。

题目:FlowAware:一种支持AI for Science任务的模型分布式自动并行方法

作者:曾艳、吴宝福、易广政、黄成创、邱扬、陈越、万健、胡帆、金思聪、梁迦隽、李欣

来源:数据与计算发展前沿, 2025, 7(5): 65-87.

摘要:本研究旨在解决AI for Science领域中深度学习模型分布式并行计算策略设计与实现困难等导致AI for Science任务计算低效的问题。本文提出了一种支持 AI for Science任务的模型分布式自动并行方法FlowAware。该方法基于AI for Science框架 JAX,深入分析AI for Science任务特征以及深度学习模型的算子结构和数据流特性,结合集群拓扑结构,构建模型分布式并行计算策略搜索空间;在此基础上以负载均衡和通信最优化为目标,为AI模型搜索最优分布式并行计算策略。 在类似GPU的加速器集群和GPU集群上进行了对比实验,实验结果表明,相比AlpaFlowAware的吞吐量最高可提升7.8倍。FlowAwareAI for Science任务中AI模型提供了高效的分布式并行策略搜索方法,并显著加速了AI模型的计算性能。

题目:数字社会的技术基础设施、运行逻辑和治理现代化

作者:韩海庭、孙茜

来源:数据与计算发展前沿, 2025, 7(5): 173-183.

摘要:。在全球经济衰退、技术竞争加剧的大背景下,数字社会技术基础设施的建设和发展不仅对于对冲我国经济下行风险至关重要,还被寄希望于促进新质生产力的开发和领导未来网络空间秩序的建设,针对数字社会的创新治理也变得尤为重要。在数字化转型背景下,本研究从身份、价值和治理对象三个方面提出创新数字社会治理的新模式。用网络标识符取代公民行为者,用分布式匿名通信取代物理交易活动,用算法、协议和代码取代合同、道德和法律等传统社会制度,将是未来数字社会的重要标志,从而使社会治理变得可衡量和可模拟。在此基础上,提出了数字身份治理、分布式账本价值管理和结果导向价值治理三种创新型社会治理新模式,为今后的研究提供一定的理论参考和研究思路。


附件: