人工智能正在重塑各行各业，掌握AI技术的人才成为市场上竞相争夺的"香饽饽"。据2024年多个行业人才需求报告显示，人工智能相关岗位的年需求量增长 率超过40%，平均薪资水平比传统计算机专业高出30%以上。 当清晨的城市交通被智能算法优化，手机上的应用通过你的语音指令预测需求，在医生做出诊断前，AI系统已经分析了海量影像资料，这正是人工智能 （AI）专业日渐火爆的时代背景。根据国家"十四五"规划和2035年远景目标纲要，我国正在加快数字化发展，特别是"人工智能+"行动已上升为国家战略。 人工智能正从科幻概念转变为推动社会发展的核心动力，融入经济、医疗、教育、交通等各行各业。 01.AI专业成为热点 AI专业的核心吸引力在于其强大的跨学科融合能力，一个AI专业的学生同时会接触到数学、计算机、数据科学甚至心理学、语言学等多领域知识。 在西安科技大学高新学院，人工智能课程早已将前沿的华为昇腾系列产品认知、华为云EI平台应用和MindSpore框架实践等内容纳入教学体系。这些课程不 仅是理论学习，更重要的是实践能力的培养。 02.西安科技大学高新学院的教育实践 在众多的陕西高校中，西安科技大学高新学院 走出了一条特色鲜 ...

Core Insights - Huawei's Developer Conference (HDC 2025) showcased the latest advancements in technology and strategic plans for the HarmonyOS ecosystem [1][3] - HarmonyOS 6 is set to enter Beta testing for developers, marking a significant update just 8 months after the release of HarmonyOS 5 [3] - The conference highlighted technological breakthroughs and ecosystem progress within HarmonyOS [3][4] Technological Breakthroughs - HarmonyOS 6 leverages advanced technologies such as Harmony AI, Pangu model/MindSpore framework, Ark Engine, and Star Shield security architecture to enhance connectivity, intelligence, security, and user experience [4] - The system supports over 50 apps for seamless sharing and collaboration, significantly improving work efficiency [6] - With over 4,000 applications integrated through AI components and 470 services connected to the intent framework, HarmonyOS aims to create a comprehensive intelligent interaction ecosystem [6] Developer and Ecosystem Innovation - The HarmonyOS ecosystem is designed to break down device barriers, enabling intelligent interconnectivity across various applications, from office software to smart home devices [7] - Over 8 million developers are contributing to the HarmonyOS ecosystem, with more than 30,000 applications and services actively being developed and updated [9] - Huawei organized various developer activities, including the 2025 HarmonyOS Innovation Competition, to stimulate creativity and innovation within the ecosystem [11] Developer Engagement and Training - The conference featured events like the Geek Marathon and CodeLabs training camp, providing developers with hands-on experience and training in advanced technologies [12] - The training camp focuses on practical applications of Huawei's cloud products, enhancing developers' skills and project experience [12] Ecosystem Growth and User Experience - The open-source HarmonyOS has accumulated over 1.3 billion lines of code and nearly 9,000 community contributors, fostering innovation across multiple industries [12] - More than 1 billion devices are now connected to the HarmonyOS ecosystem, facilitating seamless interconnectivity and enhancing user experience [14] - The advancements in the HarmonyOS ecosystem reflect its strong appeal and ability to attract developers, partners, and users, promoting a collaborative and innovative environment [14]

Core Viewpoint - The founder of Huawei, Ren Zhengfei, suggests that the Chinese chip industry can break through its current challenges through "stacking and clustering" technology paths, achieving performance comparable to the world's leading levels in computing [2] Group 1: Technological Innovations - China lags behind the US by one generation in single-chip processes, but innovative strategies like "mathematics compensating for physics" and "non-Moore compensating for Moore" can lead to breakthroughs in system-level performance [2] - Huawei's Ascend chip, while not on par with the leading 3nm chips in process technology, utilizes self-developed CCE communication protocols to create efficient clusters, supporting the training of the Pangu large model, achieving computational power comparable to top GPUs [2] - The success of Google's TPU cluster, which trained a 540 billion parameter model, demonstrates the scale effect of cluster computing in AI [2] Group 2: Algorithm Optimization - Huawei's breakthroughs in algorithm optimization are crucial, with the "mathematics compensating for physics" concept reflected in techniques like sparse computing, model quantization, and pruning [4] - The MindSpore framework reduces AI training computational demands by over 30% through dynamic optimization and low-precision computing, showcasing effective software and hardware collaboration [4] - The Tianjin Port's unmanned terminal exemplifies this, where hundreds of Ascend chips form a computing cluster to process massive sensor data in real-time, significantly enhancing operational efficiency [4] Group 3: Chiplet Technology - The "non-Moore" path emphasizes Chiplet technology, which breaks down complex large chips into smaller, functionally distinct chiplets, achieving high-density integration through 2.5D/3D packaging [4] - Huawei's Ascend chip employs a Chiplet architecture, where core computing units pursue advanced processes while I/O, analog, and storage modules use mature processes, achieving performance that rivals or exceeds that of single advanced process large chips [4] - This modular design strategy is similar to AMD's Zen architecture, which has gained significant market share in servers through efficient interconnect technology [4] Group 4: Research and Collaboration - The fundamental bottleneck of China's chip industry lies in basic theoretical research, with Huawei investing 180 billion yuan annually in R&D, including 60 billion yuan for foundational research without performance metrics [6] - Huawei's contributions to game theory, such as the "prophet inequality," address optimization problems, with related papers published at the 65th Computer Science Foundations Symposium [6] - Open collaboration and talent strategies are essential for breakthroughs, with Huawei integrating resources through partnerships with open-source communities and attracting top global talent through initiatives like the "Genius Youth" program [6] Group 5: Industry Outlook - Ren Zhengfei's statements instill confidence in the Chinese chip industry, which has seen numerous companies achieve breakthroughs in mid-to-low-end sectors, particularly in compound semiconductors [6] - Huawei's "stacking and clustering" strategy not only provides a technical pathway for itself but also offers a model for the industry [6] - With ongoing investments in foundational research and the improvement of the industrial ecosystem, the Chinese chip industry is expected to achieve greater autonomy and control in more fields [6]

Core Insights - Nanjing Artificial Intelligence Computing Center focuses on integrating industry, academia, and research to foster AI talent and innovation in the Yangtze River Delta region [1][4] - The center has established deep collaborations with over 10 universities and research institutions, training more than 5,000 AI professionals to support industrial upgrades [1][3] Group 1: Research and Innovation - The center actively participates in provincial research plans, successfully applying for two projects in remote sensing and education for the 2024 Jiangsu Provincial Technology Department's frontier technology R&D program [2] - In agricultural research, the center collaborates with Nanjing Agricultural University on a "soybean breeding large model" that utilizes AI to analyze genetic data, shortening breeding cycles and providing efficient tools for seed industry innovation [2] Group 2: Talent Development - To address the shortage of AI talent, the center has developed a comprehensive training system that includes "training + practice + connection" [3] - The center has conducted 12 "Ascend University Tour" lectures in collaboration with Southeast University and Nanjing University of Aeronautics and Astronautics, reaching over 5,000 students and covering practical topics such as MindSpore framework application and large model training [3] Group 3: Ecosystem and Collaboration - The center serves as a public platform, hosting nearly 20 events such as "Experience AI Charm" and Ascend Technology Forums, attracting participation from enterprises and investment institutions [4] - This has facilitated the innovative development of over 10 large model companies in Nanjing, forming an innovation ecosystem that combines computing power platforms, research institutions, and enterprises [4] - The center aims to create a virtuous cycle of technological innovation and industrial development, reinforcing the foundation for autonomous AI development in the city [4]