Results 21 to 30 of about 5,582,534 (160)
智能科学与技术学报, 2020 深度强化学习主要被用来处理感知-决策问题，已经成为人工智能领域重要的研究分支。概述了基于值函数和策略梯度的两类深度强化学习算法，详细阐述了深度Q网络、深度策略梯度及相关改进算法的原理，并综述了深度强化学习在视频游戏、导航、多智能体协作以及推荐系统等领域的应用研究进展。最后，对深度强化学习的算法和应用进行展望，针对一些未来的研究方向和研究热点给出了建议。
刘朝阳, 穆朝絮, 孙长银
doaj
Resource Substitution: Local Governments' Issue Attention and Policy Reinvention
Policy Studies Journal, EarlyView.ABSTRACT Research on policy reinvention, where a government modifies rather than replicates policy innovations introduced by other governments, has progressed significantly, but how it is influenced by the government's issue attention has not been thoroughly examined.
Yingxin Zhang, Yixue Yao, Kaifeng Yang
wiley +1 more source
, 2019 《2019年诺贝尔化学奖的电化学解读》专辑导读 2019年诺贝尔化学奖颁给了美国德克萨斯大学奥斯汀分校的约翰·古迪纳夫（John B. Goodenough）、美国纽约州立大学宾汉姆顿分校的斯坦利·惠廷厄姆（M. Stanley Whittingham）和日本旭化成公司的吉野彰（Akira Yoshino）三位科学家以表彰他们对锂电池研发的卓越贡献.
《电化学》编辑部
core
基于强化自组织映射和径向基神经网络的短期负荷预测 [PDF]
全球能源互联网, 2019 径向基（radial basis function，RBF）神经网络因其泛化能力强、收敛速度快的特点广泛应用于负荷预测。但传统采用K-means和自组织映射（self-organizing map，SOM）训练RBF径向基中心的方法因其全局搜索能力偏弱，仍然存在容易陷入局部最优解的问题，严重制约了RBF预测精度的提高。针对此问题，提出了一种基于强化学习（reinforcement learning，RL）改进的RBF短期负荷预测方法。强化学习通过环境的反馈不断完善搜索策略，具有非常突出的全局搜索能力 ...
黄乾, 马开刚, 韦善阳, 黎静华
doaj +1 more source
Pollution and Haunted Eco‐Histories: Silent Hill and Silent Horrors in Centralia, Pennsylvania
Popular Culture Review, Volume 37, Issue 2, December 2026.ABSTRACT Centralia is a former mining town and borough in Northeastern Pennsylvania, whose population has been steadily declining since the Mine Fire of 1962—a fire that still burns today. The “horrors” of living in Centralia are reflected in horror media set in postindustrial mining towns like Silent Hill.
Maz Jardon
wiley +1 more source
World Water Policy, Volume 12, Issue 4, November 2026.ABSTRACT This article examines how renewable‐energy expansion is changing water–energy bargaining among Iran, Iraq, and Türkiye in the Tigris–Euphrates basin. Much of the existing literature on transboundary water politics focuses on territorial control, upstream–downstream power asymmetries, legal rules, and hydro‐hegemony but gives less attention to ...
Umud Shokri
wiley +1 more source
智能科学与技术学报, 2021 针对多光谱行人检测系统存在特征融合质量低、模型超参数多且锚框匹配算法复杂等问题，提出了一种基于差分特征注意力机制的无锚框多光谱行人检测算法。该算法首先采用差分特征感知融合方法挖掘多模态特征间的互补信息来优化通道特征；然后利用具有高效无锚框机制的CenterNet检测框架大大降低了模型计算复杂度，从而提升检测速度；最后引入差分特征注意力机制，改善特征融合质量，进一步提升检测精度。在KAIST、CVC14和FLIR这3个公开数据集上的实验结果表明，提出的算法和其他先进方法相比，能够同时有效提升检测精度和速度，
沈继锋 +4 more
doaj
Light‐Synergistic Enhanced Synaptic Devices Toward Superadditive In‐Sensor Computing
Rare Metals, Volume 45, Issue 10, October 2026.ABSTRACT Conventional neuromorphic devices face challenges in achieving efficient integration of light sensing and synaptic computation because of structural and material limitations. In this work, we developed an indium‐tin‐zinc oxide/transparent gelatin electrolyte synaptic transistor incorporating rare metal indium.
Wei Zhong +16 more
wiley +1 more source
Enhanced Corrosion Resistance in an Additively Manufactured Copper Alloy via Suppressed Segregation
Rare Metals, Volume 45, Issue 10, October 2026.ABSTRACT This study elucidates the corrosion behavior of C18150 Cu–Cr–Zr alloy fabricated by laser powder bed fusion (LPBF) in comparison with a conventional wrought baseline representative of industrial Cu–Cr–Zr plate products, highlighting the superior corrosion resistance of the additively manufactured condition.
Yang Qi +4 more
wiley +1 more source
智能科学与技术学报, 2021 旨在讨论智能绿色时代下港口工业的现状需求、体系结构、新形态和其关键平台技术等问题。首先介绍了智慧绿色时代下港口工业的现状和需求，并详细描述了平行港口工业体系结构与新形态；然后讨论了智慧绿色时代下港口工业中的工业智联网、知识自动化和人机混合增强智能等新一代技术；最后讨论了平行港口在智慧绿色时代中的具体应用场景和案例，重点围绕港口船舶管理系统进行分析。
吴宇震 +4 more
doaj