Core Insights - The research team from Tianjin University has developed an innovative "atomic extraction" strategy that significantly enhances the utilization of platinum atoms in catalytic reactions, achieving nearly 100% efficiency [1][2] - This advancement addresses the low atomic utilization rates of traditional catalysts, particularly in the critical process of propane dehydrogenation for propylene production, which is essential for various industries [1][2] Group 1: Technological Innovation - The "atomic extraction" technology allows for the selective attraction of platinum atoms to the surface of catalyst particles, thereby increasing their availability for catalytic reactions [1] - This method contrasts with traditional catalysts, where platinum atoms tend to aggregate into larger particles, resulting in many atoms being buried and inactive [1] Group 2: Economic Impact - The new catalyst technology reduces the amount of catalyst needed from 1 ton to just 200 jin (approximately 100 kg), showcasing extreme atomic economy [2] - In high-temperature reactions like propane dehydrogenation, the new catalyst demonstrates superior activity and stability while reducing platinum usage by 90% compared to existing commercial catalysts [2] Group 3: Industry Relevance - Propylene is one of the most produced chemical products globally, with China projected to account for one-third of global propylene production by 2024, generating over 600 billion RMB [1] - Approximately two-thirds of propane dehydrogenation processes rely on precious metal catalysts, highlighting the importance of improving catalyst efficiency for sustainable industry development [1]

长着"眼睛"，能自主行动，并进行团队合作……在近日举行的第二十五届中国国际工业博览会上，工业 场景中的机器人成为吸引观众的明星。在人工智能技术的加持下，机器人不再只是按部就班地执行程序 操作，而是拥有了自主性，开始以柔性能力处理生产过程中各项无法事先安排的情况，成为人类生产者 们能干的"同事"。 AI技术也能帮助机器人变得更加绿色——通过算法优化运动轨迹节能，或是在待机时将机器人调整至 低能耗状态。这些看似微小的优化，在规模化部署中将汇聚成推动制造业可持续发展的重要动力。 在实际生产中，机器人必须能在复杂环境中行走自如。科技日报记者在现场看到，上海ABB工程有限 公司（以下简称"ABB"）展台上，机器人通过"基于外部引导运动"功能，能在生产环境中快速、准确地 跟踪和优化运动路径，灵活避开障碍物，精准完成各种复杂操作。ABB全球首发的OmniCore EyeMotion视觉系统如同机器人的"眼睛"一样，可让机器人使用任意第三方相机或传感器感知周围环 境。 （文章来源：科技日报） AI+机器人的生态互补进一步提升了机器人工业场景的"含智量"。在节卡机器人股份有限公司展台， AI+工业智能机器人操作系统——JAKA ...

Core Viewpoint - The Karachal Fluorspar Mine, designed by China ENFI Engineering Corporation, has commenced production, marking it as the largest primary fluorspar mining and selection base in China and Asia, with a resource endowment that accounts for approximately 15% of the country's proven fluorspar reserves [1][2]. Group 1: Project Overview - The Karachal Fluorspar Mine has a production capacity of 1.2 million tons per year and is recognized for its superior resource endowment [1]. - The mine has achieved significant breakthroughs in the efficient selection of fluorspar from difficult-to-process calcite-type ores, overcoming major technical challenges such as the suppression of calcite, low-temperature selection, and coarse particle flotation [1]. - The project has designed a fluorspar recovery rate of over 80% and a fluorspar concentrate grade of 97%, establishing an engineering technology system for the efficient selection of calcite-type fluorspar [1]. Group 2: Technological Innovations - The Karachal Fluorspar Mine is the first non-coal underground mine in China to fully adopt non-explosive mechanical continuous mining technology during the design phase [2]. - This new mining process allows for parallel continuous operations in underground recovery, which is expected to significantly enhance mining efficiency, labor productivity, and safety management levels [2]. Group 3: Environmental and Energy Efficiency - The tailings dam of the project employs an upstream construction method and has been designed to adapt to cold climate conditions [1]. - A solar heat collection and storage system has been established to provide hot water for heating and slurry warming, thereby reducing electricity load and achieving significant energy savings [1].

全国首个司法行政综合垂直大模型"鲲鹏矩阵"日前在广东省湛江市正式发布。"鲲鹏矩阵"大模型由湛江 市司法局打造，立足司法行政工作实际，构建了从底层算力支撑、数据资源整合到大模型能力建设、上 层智能应用集成的全链条智能化工作体系。相比通用大模型，"鲲鹏矩阵"更贴合司法行政工作的专业特 性，有效解决了以往人工智能应用泛而不精的问题。 今年2月，湛江市司法局成立司法行政模板探索课题组，由该局党组书记、局长罗东玲担任课题组组 长，集结业务骨干，系统梳理行政复议、社区矫正、人民调解等司法行政业务流程，精准定位业务痛 点、难点、堵点，最终形成18个覆盖全业务链的场景模块，从具体业务切入开展人工智能落地探索。 目前，基于"鲲鹏矩阵"，湛江市司法局已成功研发并应用80余个智能体，覆盖法律援助、立法普法、依 法治市等核心场景。如"鲲鹏智答"智能服务平台可为群众提供7×24小时智能法律咨询，已累计服务超 2.6万人次，90%的用户评价其解答"专业、易懂、实用"。据悉，大模型将在该市122个司法所全面铺 开，推动司法行政工作从经验驱动向数据驱动、智能驱动转变。 罗东玲表示，湛江市司法局将持续优化模型能力，拓展应用场景，大力推动科技创 ...

Core Insights - The launch of China's first application-level embodied robot for home use marks a significant advancement in the field of household service robots, integrating "embodied intelligence + professional services" [1] Group 1: Product Development - The robot was co-developed by the Institute of Automation, Chinese Academy of Sciences, and Starry Power, focusing on AI, large model semantic understanding, and multimodal human-computer interaction technologies [1] - The product aims to create an efficient home health management system, specifically targeting chronic disease management and Alzheimer's disease prevention [1] Group 2: Collaboration and Support - Starry Power collaborates with Qingdao Yinghai Hospital, a subsidiary of Haier, to ensure the product's functionality is aligned with healthcare needs, creating a closed-loop health service process [1] - The robot's development is supported by clinical data and medical solutions provided by Qingdao Yinghai Hospital [1] Group 3: Technological Capabilities - The robot features multimodal interaction capabilities, allowing it to perceive the home environment, understand voice commands, and autonomously learn to optimize service strategies [2] - It can actively collect physiological indicators and provide customized health plans, medication reminders, and rehabilitation guidance [2] Group 4: Future Applications - Future applications of the robot are expected to extend to community and professional health care institutions, contributing to an integrated smart health care ecosystem that encompasses home, community, and institutions [2]

记者从陆军军医大学新桥医院（以下简称"新桥医院"）获悉，该院联合重庆大学计算机学院，经过半年 多协同攻关，于日前成功研发出肺部感染疾病的人工智能专病大模型。这是国内首个基于诊疗策略生成 的肺部感染专病大模型。目前，该模型已部署至新桥医院呼吸病区医护工作站，进入临床试用阶段。 （文章来源：科技日报） 新桥医院呼吸与危重症医学中心主任徐智介绍，肺部感染作为一种常见却极具挑战性的疾病，每年全球 新增病例数亿，死亡人数超过百万。其病因复杂，表现多样，给临床诊断和治疗带来很大困难。对老年 人、孕产妇等特殊人群以及病情复杂的患者来说，医生在制定治疗方案时往往面临巨大挑战。 该模型聚焦肺部感染诊疗痛点，整合12893例高质量临床病例与包含43万余条问答的国内外权威医学指 南，为肺部感染疾病诊疗提供智能解决方案。在技术实现过程中，团队摒弃传统方法，结合多种训练策 略，使模型能更精准地理解医学问题，提供个性化诊疗建议。 新桥医院信息科主任陈学涛介绍，团队建立了"文本相似度+医疗关键点覆盖"的立体评测体系，该模型 在多项测试中表现出色。与当前国内外先进的通用大模型相比，这款模型在肺部感染疾病诊断准确性和 治疗方案推荐合理性上均有 ...

Core Insights - Scientists at the University of Cambridge's Cavendish Laboratory have observed a quantum effect in organic materials, previously thought to exist only in inorganic metal oxides, which allows for highly efficient conversion of light energy into electrical energy [1][2] - The research focuses on a spin-free radical organic semiconductor called P3TTM, which contains unpaired "single" electrons that exhibit unique magnetic and electronic behaviors [1] - The study reveals that when these molecules are densely packed, the "single" electrons display "Mott-Hubbard insulator" behavior, leading to alternating arrangements of electrons [1] Group 1 - The "single" electrons absorb light energy and transition to neighboring molecules, creating a chain reaction that separates positive and negative charges, resulting in a continuous electric current [2] - A solar cell was constructed using P3TTM films, achieving nearly perfect charge collection efficiency, with almost every incident photon converted into usable current [2] - Unlike traditional organic solar cells that require two materials for electron transfer, the new material can complete the entire conversion process using a single substance, potentially leading to lower-cost and lightweight solar cells [2]

Core Insights - The event held on October 16, 2025, in Guangzhou focused on the practical applications of AI, emphasizing its role as a core engine for urban governance and enterprise development, moving beyond superficial perceptions of AI technology [1][2] Group 1: AI in Urban Management - Traditional streetlight management faces challenges such as reliance on manual inspections, delayed citizen reporting, and dependence on expert experience, which AI can effectively address [2] - The team at Zhongke Zhicheng has developed an AI streetlight operation and maintenance platform, achieving a fault identification accuracy rate of 95.8%, up from 70%, and a detection rate improvement from 30% to 82.1% [2] - This innovative solution significantly reduces streetlight maintenance costs and enhances operational efficiency, providing strong support for the intelligent upgrade of urban infrastructure [2] Group 2: AI in Enterprise Management - The company introduced a localized large model combined with a localized knowledge base to facilitate data flow across different systems, breaking down barriers to information sharing [3] - By encouraging employees to contribute to a corporate knowledge repository, the solution enables precise knowledge retrieval, intelligent Q&A, and personalized report generation [3] - The rapid development of AI necessitates a balanced approach that embraces its potential while addressing ethical considerations and focusing on human welfare [3]

Core Insights - A research team led by Professor Tian Hui from Peking University has successfully detected millisecond-level radio burst signals from stellar sunspot regions using the FAST telescope, providing a new observational method for studying stellar magnetic activity and its origins [1] Group 1: Research Findings - The detection of radio signals indicates that magnetic activity in stellar sunspot regions can accelerate electrons to extremely high energies, which then produce unique radio emissions as they spiral in the magnetic field [1] - This breakthrough allows for direct analysis of the small-scale magnetic field structure on stellar surfaces, enhancing the understanding of complex stellar magnetic fields [1] Group 2: Technological Advancements - The success of this research is attributed to the high sensitivity and resolution of the FAST telescope, which has improved the time resolution of stellar radio observations to the sub-millisecond level, capturing minute variations in stellar radio emissions [1] - Currently, there are very few devices worldwide that can match the capabilities of FAST in this regard [1]

Core Viewpoint - The research team led by Professor Tian Hui from Peking University has successfully detected millisecond-level radio burst signals from the sunspot region of a star using the FAST telescope, providing a new observational method for studying stellar magnetic activity and its origins [1][4]. Group 1: Stellar Magnetic Activity - Sunspots are localized regions of strong magnetic fields on the surface of stars, and their magnetic activity can disrupt the space environment around Earth, affecting satellite operations and communication systems [4]. - Traditional methods for measuring stellar magnetic fields have struggled to capture small-scale magnetic structures, often only providing information on large-scale magnetic fields [4]. - The detection of a unique radio burst from the active red dwarf star AD Leo, with a frequency drift rate of approximately 8 GHz per second, indicates that the radiation source is from a high magnetic field region very close to the star's surface [4][5]. Group 2: Implications of the Discovery - The findings suggest that magnetic activity in stellar sunspot regions can accelerate electrons to high energies, producing unique radio emissions that allow for direct analysis of small-scale magnetic field structures on stars [5]. - The success of this research is attributed to the high sensitivity and resolution of the FAST telescope, which has improved time resolution to the sub-millisecond level, enabling the capture of minute variations in stellar radio emissions [5]. - This research opens new avenues for studying magnetic activity in late-type stars and exploring interactions between stars and planets, potentially aiding in the search for habitable exoplanets [5].