明敏 发自 凹非寺 量子位 | 公众号 QbitAI 华为云算力再迎重大突破！ 刚刚落幕的华为全联接大会2025，一系列新进展发布—— 这距离CloudMatrix384超节点2025年4月正式发布仅半年，期间其 能力持续进化 ： 现阶段， AI行业内依旧被算力焦虑笼罩 。硅谷大厂近期在算力、芯片领域动作频频： OpenAI一边和博通自研AI芯片，一边向甲骨文抛出3000亿美元买算力；马斯克百天建成万卡超算集群，还计划向百万卡规模冲击，同时悄悄 布局芯片；Meta、AWS等企业也在积极获取更多算力资源……但算力的发展并非一蹴而就，它需要在单点技术上极致突破，还涉及芯片、硬 件、架构、软件、网络、能源乃至整个产业生态的协同演进。 放眼全球，能够输出澎湃算力的供应商，都离不开十数年、数十年的沉淀积累。 华为云作为其中一员，探索路径因所处产业阶段而显得尤为深刻：不仅需要在技术"无人区"重新定义算力运行规则；还需把握AI发展时机，通 过快速迭代响应产业海量需求。一步步成长为今天的"算力黑土地"。 AI算力云服务升级， 基于华为云刚刚发布的最新AI服务器规划， CloudMatrix的云上超节点规格将从384卡升级到未 ...

Core Viewpoint - The article discusses the integration of artificial intelligence (AI) into scientific research, highlighting Beijing's role as a leader in this transformation through the AI for Science (AI4S) initiative, which aims to enhance research efficiency and innovation [1][2][6]. Group 1: AI4S Development and Impact - AI4S is recognized as a new paradigm that accelerates scientific research, moving from theoretical concepts to practical applications, with a global consensus on its importance [2][4]. - The establishment of the Beijing Scientific Intelligence Research Institute in 2021 marks a significant step in combining AI with scientific research, aiming to create a new research infrastructure [3][5]. - The DeepFlame Rocket software exemplifies the transition from research to commercial applications in AI4S, significantly reducing the time and cost of rocket engine development [4][5]. Group 2: Infrastructure and Ecosystem - Beijing has developed a comprehensive innovation and industrial chain in AI, producing significant original achievements such as a large atomic model covering over 90 elements and a new generation of research knowledge databases [5][9]. - The "Bohler Research Space Station" serves as a foundational infrastructure for AI4S, providing tools for literature management, interdisciplinary knowledge discovery, and experimental calculations [7][8]. - The OpenLAM project aims to create a large atomic model to facilitate micro-scale design in various industries, demonstrating the collaborative efforts in building research infrastructure [10][11]. Group 3: Future Plans and Goals - The "Action Plan" outlines Beijing's strategy to establish a scientific foundation model and high-quality scientific databases by 2027, targeting over 10 million users and fostering an open-source ecosystem [12][13]. - The plan emphasizes the importance of collaboration among research institutions and enterprises to enhance the AI4S innovation ecosystem and drive significant scientific advancements [12][13].

Core Viewpoint - China is accelerating its journey towards becoming a technology powerhouse, with significant advancements in various scientific and technological fields during the "14th Five-Year Plan" period [1]. Group 1: Technological Innovations - Major achievements include the operation of the world's first fourth-generation nuclear power plant, the successful return of the Chang'e 6 lunar mission, and the completion of the Chinese space station [1]. - In deep space exploration, the Chang'e 6 mission revealed volcanic activity on the moon's far side dating back 2.8 billion years [3]. - In quantum technology, China achieved a world record by preparing and verifying a cluster state of 51 superconducting qubits, surpassing all other quantum systems [3]. - The number of China's hot research papers continues to grow, maintaining the top global ranking as of July 2024 [3]. Group 2: Economic Impact of Innovation - R&D investment in China reached a new high, with a nearly 50% increase compared to the end of the "13th Five-Year Plan," amounting to an additional 1.2 trillion yuan [6]. - The "new economy" represented by "Artificial Intelligence+" generated over 24 trillion yuan in added value in 2024, equivalent to the combined GDP of Beijing, Shanghai, and Guangdong [6]. Group 3: Challenges and Strategic Responses - Despite significant progress, challenges remain in integrating the innovation chain with the industrial chain, and improving the efficiency of converting innovation results into productive forces [9]. - The Chinese government has issued important policies to promote the application of artificial intelligence, signaling a commitment to addressing uncertainties in technology and external environments [9]. - The transition of labor skills is increasingly pressing, necessitating a focus on releasing innovation energy where it is most needed [9].

Group 1 - The core viewpoint emphasizes China's significant advancements in technology and innovation during the "14th Five-Year Plan" period, showcasing numerous groundbreaking achievements in various fields [1][2][3] - China's focus on foundational research has led to major breakthroughs, such as the successful return of lunar samples by Chang'e 6 and advancements in quantum technology, indicating a strong commitment to high-level scientific independence [2][3] - The integration of technology and industry is accelerating, with a notable increase in R&D investment, which has grown nearly 50% compared to the end of the "13th Five-Year Plan," amounting to an additional 1.2 trillion yuan [4][5] Group 2 - The launch of the first commercial version of a domestic operating system based on the 6.6 kernel marks a significant step in supporting AI applications and digital transformation across industries [4] - The "New Economy" sector, represented by "Artificial Intelligence +," has seen its added value exceed 24 trillion yuan in 2024, equivalent to the combined GDP of Beijing, Shanghai, and Guangdong [4] - The recent World Artificial Intelligence Conference highlighted the transformative impact of AI on scientific methodologies, with new models being developed to enhance researchers' creativity [5][6] Group 3 - Challenges remain in the integration of innovation and industry, with a need to improve the efficiency of converting innovative results into productive forces and addressing the skills transition of the workforce [6] - The Chinese government has issued important directives to promote the development of AI applications, signaling a commitment to navigating the uncertainties posed by rapid technological changes [6] - China's transition from a major technological power to a strong technological nation is supported by a solid foundation, with an emphasis on maintaining confidence in innovation and striving for global technological leadership [6]

Group 1 - The core viewpoint emphasizes China's significant advancements in technology and innovation during the "14th Five-Year Plan" period, showcasing numerous groundbreaking achievements in various fields [1][2][3] - China's focus on foundational research has led to major breakthroughs, such as the successful return of samples from the Chang'e 6 mission and advancements in quantum technology, indicating a strong commitment to scientific exploration [2][3] - The integration of technology and industry is accelerating, with a notable increase in R&D investment, which has grown nearly 50% compared to the end of the "13th Five-Year Plan," reaching an additional 1.2 trillion yuan [4] Group 2 - The launch of the first commercial version of a domestic operating system based on the 6.6 kernel marks a significant step in supporting AI applications and digital transformation across industries [4] - The "New Economy" sector, represented by "Artificial Intelligence +," is rapidly developing, with its added value exceeding 24 trillion yuan in 2024, equivalent to the combined GDP of Beijing, Shanghai, and Guangdong [4] - The need for innovation to address existing challenges is highlighted, with a focus on enhancing the efficiency of converting innovative results into productive forces and addressing the skills transition of the workforce [6]

Group 1 - China's innovation has reached unprecedented importance during the 14th Five-Year Plan, with significant achievements in various sectors, including space exploration and quantum technology [1][2] - The successful return of samples from the Chang'e 6 mission reveals volcanic activity on the moon's far side, showcasing advancements in deep space exploration [1] - The number of China's hot research papers continues to grow, maintaining the world's leading position as of July 2024 [1] Group 2 - The release of a commercial version of a desktop and server operating system based on the 6.6 kernel marks a significant step in the integration of technology and industry, providing a new foundation for AI applications and digital transformation [2] - R&D investment in China has reached a new high, with a nearly 50% increase compared to the end of the 13th Five-Year Plan, amounting to an additional 1.2 trillion yuan [2] Group 3 - The "New Three" economy, represented by new industries, new business formats, and new models, is accelerating, with a total value exceeding 24 trillion yuan in 2024 [3] - The World Artificial Intelligence Conference highlighted the transformative impact of AI on scientific methodologies, with new models being developed to enhance researchers' creativity [3] Group 4 - The State Council's recent implementation of the "Artificial Intelligence+" action plan signals a commitment to advancing AI applications amid rapidly changing technological and external environments [4] - China's transition from a major technological power to a strong technological nation is supported by a solid foundation, emphasizing the need for innovation confidence and contributions to global scientific progress [4]

Core Viewpoint - The article introduces a new "neuro-symbolic" hybrid planner developed by the Chinese Academy of Sciences, which significantly enhances the efficiency and precision of scientific research planning compared to traditional methods [1][5]. Group 1: Mechanism and Features - The hybrid planner integrates the advantages of both neural planning systems and symbolic planning systems, leading to improved expressiveness, adaptability, generalization, and interpretability [3][11]. - It employs a closed-loop feedback mechanism inspired by human motor learning, enhancing the planner's ability to detect and correct errors dynamically [10][6]. - The system features a self-control mechanism that allows the planner to determine when to receive feedback, optimizing the frequency of feedback and reducing dependency on it [18][21]. Group 2: Performance Evaluation - The hybrid planner was evaluated against eight representative planning tasks in the International Planning Competition (IPC), showing an average coverage rate of 70.81%, which is significantly higher than other comparative planners [23][25]. - In a comparison with OpenAI's o1 model on the PlanBench dataset, the hybrid planner achieved 100% coverage and significantly reduced average planning time, demonstrating its superior efficiency and effectiveness [26][25].

Core Insights - The "AI Scientist Agent Platform" has been officially launched, aiming to create an accessible research assistant for scientists [1][2]. Group 1: AI Scientist Agent Platform - The platform is developed based on the "Panshi Scientific Foundation Model" by the Chinese Academy of Sciences, focusing on literature research, scientific reading, and frontier tracking [1][2]. - It features capabilities such as scientific literature extraction, knowledge representation, reasoning, and tool orchestration, making it a multi-disciplinary intelligent research assistant [2]. - The platform utilizes over 200 million real-time updated scientific literature resources to identify cutting-edge results and generate multi-dimensional research reports [2]. Group 2: Applications in Research - The AI Scientist Agent is applicable across various user groups, including universities, research institutes, and technology companies, providing targeted intelligent support for different research stages [3]. - For university faculty and students, it assists in course teaching, topic selection, literature research, and writing, significantly enhancing research training efficiency and quality [3]. - For technology company R&D teams, it accelerates technology foresight research, patent analysis, and interdisciplinary data integration, improving R&D and decision-making efficiency [3]. Group 3: Decision Intelligence Operating System (DIOS) - The DIOS decision intelligence operating system was also launched, aiming to integrate the entire chain from computing power, data, models, to decision applications [4]. - It provides a comprehensive AI service chain, significantly reducing deployment costs and time while ensuring safety and reliability [4]. - The platform combines the "Panshi Scientific Foundation Model" with a self-developed decision model, supporting intelligent applications across multiple industries [4].

Group 1: AI Innovations in Production and Daily Life - AI is leading a new technological revolution, enhancing productivity and creativity in various fields, such as video generation and 3D modeling [2][3] - The "Keling AI" platform allows users to create videos from text inputs, significantly improving creative efficiency and freeing up creators' productivity [3] - Smart glasses showcased at the conference enable users to interact and place orders in different languages, demonstrating practical applications of AI in everyday scenarios [5] Group 2: AI's Impact on Scientific Research - AI is reshaping the foundational logic of scientific research, accelerating discoveries and providing new opportunities to solve significant scientific challenges [6] - Notable advancements include AI's ability to predict protein folding structures and assist in various scientific fields such as quantum computing and cancer detection [6][7] - The "Panshi Scientific Foundation Model," developed by the Chinese Academy of Sciences, aims to enhance scientific tasks through deep understanding of scientific data and literature [6][7] Group 3: Global Cooperation and Governance in AI - The development of AI is recognized as a global issue that requires collaborative governance and innovation to maximize its benefits across industries [8] - The "Global Governance Action Plan for Artificial Intelligence" was introduced, emphasizing the need for cooperation in AI innovation, safety governance, and equitable development [8] - China is actively working to bridge the intelligence gap and promote shared benefits through AI applications in education, agriculture, and healthcare across various regions [9]

Core Insights - China has released 1,509 large models, leading globally with 3,755 total large models published, indicating a rapid advancement in foundational large models and potential breakthroughs in AI technology [1][11] - The "Panshi Scientific Foundation Large Model," developed by a team from the Chinese Academy of Sciences, was officially launched, showcasing capabilities in scientific data understanding and support for scientific tasks [1][3] - The AI conference featured over 3,000 cutting-edge exhibits, including more than 40 large models, highlighting the accelerated application of foundational large models across various sectors such as healthcare, consumer goods, and education [5][11] Industry Developments - The rapid evolution of large models is evident in the education sector, with applications like a scanning pen that provides correct answers and step-by-step explanations, enhancing user interaction [7] - Open-source has become a preferred option for many large model products, with numerous companies debuting their latest models at the AI conference [9] - The global AI industry comprises over 35,000 companies, with more than 5,100 in China, representing approximately 15% of the global market, indicating a growing industrial scale and a favorable development landscape for AI in China [11]