Workflow
通信技术
icon
Search documents
写好数字金融大文章!杭州银行携20余项金融科技成果亮相2025中国国际金融展
Zhong Guo Jing Ji Wang· 2025-06-20 11:19
Core Insights - The 2025 China International Financial Expo was held in Shanghai, showcasing over 400 financial institutions and technology companies, with Hangzhou Bank being a prominent participant [1] - Hangzhou Bank emphasizes serving the real economy and has showcased over 20 digital financial solutions, reflecting its digital transformation achievements [3] Group 1: Technological Transformation - Hangzhou Bank is committed to self-controlled research and development, investing in cloud-native applications, intelligent observability systems, and big data platforms to support rapid business growth [4] - The bank successfully launched a new generation of distributed, cloud-native core business systems in November 2023, which won the "2023 Financial Technology Development Award" from the People's Bank of China [4] - Significant advancements in AI have been made, with AI technology being integrated into various operational areas, transforming tools into productivity outputs [4] Group 2: Innovative Financial Services - The bank collaborated with Huawei to create the Harmony Smart Hall ecosystem, enhancing service experiences through a standardized technical framework and a comprehensive application ecosystem [5] - The "Four Connections and Three Easies" comprehensive financial service ecosystem was introduced, integrating seven core products to meet the full lifecycle financial needs of government and enterprise clients [7] - The "Xinyibao" platform serves as an ideal partner for small and medium-sized enterprises in their digital transformation, offering various integrated services [7] Group 3: Future Focus - Hangzhou Bank aims to leverage opportunities in financial digital transformation and open innovation, focusing on enhancing core financial technology capabilities [7]
“中国—中亚机制为多边合作提供新范式”
Ren Min Ri Bao· 2025-06-18 21:53
"在地缘政治动荡和全球治理体系深刻变革的背景下,中国—中亚机制为多边合作提供新范式,为地区 国家经济社会发展带来新机遇。"乌兹别克斯坦上海合作组织民间外交中心主任萨比罗夫日前在接受本 报记者采访时表示,第二届中国—中亚峰会极大促进了中亚国家和中国相互尊重、长期互利的合作关 系。 萨比罗夫认为,乌兹别克斯坦同中国建交30多年来,特别是随着高质量共建"一带一路"深入推进,两国 关系的大树愈发根深蒂固、枝繁叶茂,各领域合作取得丰硕成果。"乌兹别克斯坦与中国交往历史源远 流长,古丝绸之路见证了人民之间的友好情谊。"萨比罗夫表示,古丝绸之路将乌中两国紧密相连,如 今,这条历史纽带被赋予新的时代内涵,搭建起互联互通、经贸往来、人文相亲的重要桥梁,为推动世 界各国和平合作、开放包容、共同发展发挥了关键作用。 "乌中两国在纺织、建筑、化工和食品等诸多行业建立了合资企业,中企积极参与乌兹别克斯坦高速公 路、桥梁、物流中心和能源基础设施等建设;比亚迪工厂在乌兹别克斯坦落地助力绿色转型发展,华为 等中企积极参与乌兹别克斯坦数字化进程,为互联网发展和青年技术培训作出了巨大贡献……"萨比罗 夫长期关注中国企业在乌兹别克斯坦的发展,对两 ...
中国海洋发展治理成果和努力在联合国海洋大会获高度评价
Xin Hua Wang· 2025-06-12 22:23
Core Viewpoint - The United Nations Ocean Conference in Nice, France, highlighted China's achievements in marine ecological protection and sustainable development, receiving high praise from global experts and officials [1][2]. Group 1: China's Initiatives and Achievements - China has established marine protected areas and wetland management zones, invested in ecological restoration, and implemented effective patrols to enhance marine biodiversity [2]. - The "Blue Partnership" initiative showcases China's commitment to sharing knowledge and technology with the international community, aiming to assist countries in capacity building for marine protection [1][3]. - The COAST toolkit, developed by China's First Institute of Oceanography, integrates marine big data, advanced numerical models, and AI technology to provide visual products and services for marine disaster management, ecological health, and fisheries resources [2]. Group 2: International Collaboration and Recognition - Experts from various countries, including Egypt and the United Nations, acknowledged the COAST toolkit's potential to help global southern nations better understand and assess their marine resources [3][4]. - China's efforts in promoting environmentally friendly regulations for deep-sea mining were commended, emphasizing its role in multilateralism and adherence to international marine rules [4]. - The International Council for Science's marine expert group highlighted China's leadership in projects under the "UN Decade of Ocean Science for Sustainable Development" (2021-2030), reinforcing its commitment to knowledge sharing and exploration [4].
武汉共青团推出安居服务体系 “青筑链”助青年“安一个家”
Core Points - The article highlights Wuhan's initiatives to support young talent through housing services, showcasing the city's commitment to attracting and retaining young professionals [1][8] - The "Youth Housing Chain" service system has been established to provide various housing options and support for young people, including free accommodation and rental discounts [1][4] - The city has implemented policies that offer significant rental reductions and housing subsidies for recent graduates and young professionals [4][7] Group 1: Housing Initiatives - Wuhan has created youth hostels and talent apartments, providing free accommodation for up to 7 days for young job seekers [2][4] - The "One Number Station" is the first youth talent apartment project in Wuhan, offering both short-term and discounted long-term rental services [2][3] - The city has integrated 29 rental housing projects to form "Youth Communities," providing 19,766 housing units for 24,543 young residents [5][6] Group 2: Support Services - The youth hostels offer additional services such as employment packages, one-on-one customer service, and recreational activities to enhance the experience for young residents [2][3] - Community spaces have been developed to host various social and educational activities, fostering connections among young people [6] - The city has established night schools and study rooms to support continuous learning and skill development for young residents [6] Group 3: Policy Enhancements - Recent policy upgrades include rental exemptions for doctoral and master's graduates, as well as discounts for university students [4][8] - The "Spring Bamboo Action" subsidy program provides financial incentives for young homebuyers, with up to 150,000 yuan for doctoral graduates [7][8] - The city has conducted outreach activities to raise awareness of housing policies, reaching over 20,000 young people [8]
从被动维权到提前防护,济南已建成商业秘密保护基地11家
Qi Lu Wan Bao Wang· 2025-06-06 04:13
优化服务供给,激发企业活力 6月5日,山东省第三届企业商业秘密保护能力提升服务月启动仪式暨创新试点工作推进会议在济南召 开。 近年来,济南市以商业秘密保护创新试点为抓手,坚持保护创新与促进发展并重、规范监管与优化服务 并举,持续提升商业秘密保护质效,切实筑牢强企护链屏障。 强化体系构建,筑牢保护根基 济南市委、市政府高度重视商业秘密保护工作,将其有效融入国家知识产权保护示范区建设,统筹考 虑、统一部署、一体推进。济南立足区域产业特色,围绕生物医药产业、化工企业、小微企业等,分行 业、分领域、分类型建立商业秘密保护指引,助力企业从"被动维权"向"提前防护"转变。目前,济南高 标准建成商业秘密保护基地11家、商业秘密保护指导站(联系点)92个,培育商业秘密保护领跑企业62 家,完善行政保护与司法保护信息互通、工作衔接机制。2021年以来,济南市中级人民法院审结侵犯商 业秘密民事诉讼案件51件,公安机关破获侵犯商业秘密案件9件,市场监管部门立案查处侵犯商业秘密 案件16件,有力维护了各类经营主体合法权益。 山东知识产权司法鉴定中心联合14家部门单位,建设全国首家商业秘密联动保护公共服务平台,为全省 企业提供宣传培训 ...
我们在海底“种”珊瑚
Core Viewpoint - The article highlights the efforts of research teams from Xiamen University and Guangxi University in coral reef restoration and conservation in China, emphasizing the importance of coral reefs as vital ecosystems and the innovative methods being employed to protect and restore them in response to climate change and human activities [4][5][6][10]. Group 1: Coral Reef Research and Restoration Efforts - Xiamen University has established a marine observation and experimental station on Dongshan Island to monitor and protect coral reefs, which are crucial for biodiversity and ecological health [5][6]. - The research team has implemented a real-time underwater monitoring system to analyze coral growth and fish activity, significantly enhancing data collection and analysis efficiency [7]. - Guangxi University has developed a comprehensive coral restoration system, including species selection, breeding, and transplantation, to address coral degradation effectively [8][9]. Group 2: Impact of Climate Change and Human Activities - Climate change, particularly rising sea temperatures, is identified as the primary cause of coral bleaching, prompting research teams to explore heat acclimatization techniques for corals [9][10]. - Human activities, such as fishing and tourism, have also negatively impacted coral reefs, leading to the implementation of protective measures like buoy deployment to prevent anchor damage [6][10]. Group 3: Community Involvement and Volunteer Efforts - The involvement of volunteers has been crucial in coral conservation efforts, with over 10,000 volunteers participating in various activities across China, particularly in Guangdong [15]. - The article emphasizes the collaborative nature of coral restoration, highlighting the contributions of local communities and volunteers in supporting scientific research and conservation initiatives [15].
一张卡干俩活,华为要把算力榨干
虎嗅APP· 2025-06-05 14:24
Core Viewpoint - The article discusses the advancements in AI, particularly focusing on Huawei's innovations in the MoE (Mixture of Experts) architecture and the introduction of RL (Reinforcement Learning) post-training techniques, which aim to enhance the efficiency and performance of large language models (LLMs) in the competitive AI landscape [1][3]. Group 1: MoE Architecture and Huawei's Innovations - The MoE model, originally proposed by Canadian scholars, has evolved significantly, with Huawei introducing the MoGE architecture that addresses inefficiencies in the traditional MoE model, leading to cost reduction and improved training and deployment [1]. - Huawei's approach emphasizes the importance of creating a collaborative ecosystem to foster the growth of the Ascend ecosystem in China [1]. Group 2: RL Post-Training Techniques - RL post-training has emerged as a critical pathway to enhance LLM performance, with models like OpenAI's o1 and DeepSeek-R1 leveraging this technique to improve reasoning capabilities in complex tasks [3][5]. - The RL post-training phase currently consumes 20% of the total computational resources, projected to rise to 50%, significantly impacting model performance and costs [3]. Group 3: Challenges in RL Post-Training - The traditional On-Policy algorithms create a "computational black hole" due to the alternating execution of training and inference tasks, leading to underutilization of resources [6][7]. - The complexity of task scheduling in large-scale clusters, exacerbated by the adoption of various parallel strategies, poses significant challenges for efficient resource utilization [8]. Group 4: Innovations in Resource Utilization - Huawei's RL Fusion technology allows a single card to handle both training and inference tasks simultaneously, effectively doubling resource utilization and throughput [9][10]. - The StaleSync mechanism enables near-asynchronous execution of tasks, achieving over 90% efficiency in horizontal scaling across CloudMatrix 384 super nodes [16][20]. Group 5: Performance Metrics and Results - The combination of RL Fusion and StaleSync has led to a significant increase in efficiency, with single-node throughput improving by 78.5% and overall performance enhancement of 1.5 times [30][31]. - StaleSync's implementation in cluster scaling shows a linear throughput increase from 35k tokens/s to 127k tokens/s as the number of super nodes increases, demonstrating its effectiveness in enhancing scalability [32]. Group 6: Conclusion - The advancements in RL post-training techniques by Huawei represent a significant leap in AI efficiency, positioning the company as a key player in the next generation of AI technology [33].
高凌信息首亏5228万营收两连降 收购欣诺通信100%股份告吹
Chang Jiang Shang Bao· 2025-06-03 23:40
Core Viewpoint - The acquisition of 100% equity in Xinnuo Communication by Gaoling Information has been terminated due to a lack of consensus among the parties involved in the transaction [1][5]. Group 1: Acquisition Details - Gaoling Information planned to acquire Xinnuo Communication for six months but announced the termination of the acquisition on June 2, 2024 [1]. - The termination was attributed to the failure of the parties to reach an agreement on the final transaction plan [1][5]. - Xinnuo Communication had submitted an IPO application in June 2023, which was accepted by the Shanghai Stock Exchange, but it was voluntarily withdrawn on June 25, 2024 [4]. Group 2: Financial Performance - Gaoling Information has experienced a decline in revenue for two consecutive years, with reported revenues of CNY 5.17 billion, CNY 3.56 billion, and CNY 2.66 billion from 2022 to 2024, reflecting year-on-year changes of 4.48%, -31.13%, and -25.23% respectively [7]. - The net profit has also decreased for three consecutive years, with figures of CNY 884.17 million, CNY 460.52 million, and a loss of CNY 522.76 million for the same period, showing year-on-year declines of 25.66%, 47.92%, and 213.52% respectively [7]. - In 2024, the company reported its first loss since going public, primarily due to reduced demand in the military telecommunications equipment sector and increased accounts receivable [7][8]. Group 3: Research and Development - Despite financial challenges, Gaoling Information maintained stable R&D investment, amounting to CNY 80.04 million in 2024, which constituted 30.03% of its revenue [1][9]. - The company has focused its R&D efforts on advanced technologies such as secure fixed-line communication, integrated communication technology, and artificial intelligence, targeting key business areas like military communication network construction [9].
国内首个!高校“可见光通信”课程落地深圳这所中学
Nan Fang Du Shi Bao· 2025-06-03 15:25
南都记者近日获悉,清华大学深圳国际研究生院"可见光通信"课程落地深圳市格致中学。作为国内首个在中学阶 段引入的可见光通信课程,这不仅是格致中学校本课程体系的一次重大飞跃,更是中学阶段网络通信教学领域的 新篇章,这一创新举措充分彰显了格致中学在"中学+大学+科研机构"协同育人模式下的科学教育实践成果。 可见光通信课程首次面向中学生开设 探索"中学+大学+企业+专业机构"办学新路径 格致中学探索"中学+大学+企业+专业机构"办学新路径,精心打造"航天科创""零一创新""人文素养"三大类百余门 校本课程,与学生社团相得益彰,为学生个性发展和多赛道成长提供支持。 除了可见光通信课程,学校早在2021年就建设大湾区第一家量子计算中心,与深圳国际量子研究院达成战略合 作,联合国际量子研究院的科研人员和深圳量旋科技有限公司科研人员共同开设中学量子校本课程,成立学生量 子社团,探索量子计算在中学的科普方式。 深圳市格致中学联合广东省无线光通信工程技术研究中心、深圳市可见光通信系统重点实验室(依托单位清华大 学深圳国际研究生院),建设"可见光通信创新实验室",面向本校初中、高中学生开设"可见光通信"科创课程。 清华大学袁克虹教 ...
训练MoE足足提速70%!华为只用了3招
量子位· 2025-06-03 06:21
训练效率不足 ,甚至 一半以上训练时间都浪费在"等待"上 。 现在,为了突破MoE的训练瓶颈,华为出手了: 构建了一套名为 Adaptive Pipe & EDPB 的优化方案,开启"上帝视角",让MoE面临"交通拥堵"的训练集群, 实现无等待流畅运行。 MoE大规模训练难题:一半以上的训练时间在等待? 实践已经表明,MoE模型训练集群的效率面临两方面挑战: 首先,是 专家并行引入了计算和通信等待 。 允中 发自 凹非寺 量子位 | 公众号 QbitAI Scaling Law之下,MoE(混合专家)如今已经成为各大模型厂商扩展模型能力的制胜法宝。 不过,在高效实现模型参数规模化的同时,MoE的 训练难题 也日益凸显: 当模型规模较大时,需要切分专家到不同设备形成并行(EP),这就引入额外All-to-All通信。 与此同时,MoE层绝大部分EP通信与计算存在时序依赖关系,一般的串行执行模式会导致大量计算单元空闲, 等待通信。 其次, 负载不均会引入计算和计算等待 。 MoE算法核心是"有能者居之",在训练过程中会出现部分热专家被频繁调用,而冷专家使用率较低的情况。 同时,真实训练数据的长度不一,不同的模型层 ...