Group 1 - The core idea of the article emphasizes the transformative impact of AI on protein research, which is crucial for understanding diseases and extending human lifespan [1][15][47] - AI is accelerating the exploration of protein structures, with predictions that human lifespan could exceed 100 years, and potentially reach 150 years [1][15] - The development of advanced AI models, such as OpenComplex2, allows for the prediction of dynamic protein structures, enhancing the understanding of protein interactions and their implications for drug development [6][8][14] Group 2 - The article highlights the importance of protein research in modern medicine, as it aids in identifying disease causes and designing precise treatment plans [3][15] - AI's role in drug development is underscored, with the potential to reduce the time and cost associated with bringing new drugs to market, breaking the "double ten law" [22][24] - The integration of AI in protein research is leading to innovative diagnostic methods, such as predicting disease risks through blood plasma protein analysis [17][19] Group 3 - The article discusses the creation of a comprehensive protein database, which supports AI models in predicting protein sequences and structures, significantly improving accuracy and efficiency [12][35] - The Venus model developed by Shanghai Jiao Tong University exemplifies the application of AI in designing functional proteins for various industrial needs [36][41] - The ongoing advancements in AI-driven protein research are expected to lead to breakthroughs in treating rare and complex diseases, enhancing overall health outcomes [47]

Group 1 - The core viewpoint of the article emphasizes the importance of understanding university admission regulations as they provide critical information for candidates regarding enrollment rules and requirements [1] - Admission regulations include key information such as distribution principles of enrollment plans, admission rules, health requirements, tuition fees, and locations of study [1] - Candidates should pay special attention to specific requirements and restrictions related to the majors they are applying for [1] Group 2 - The article highlights that various universities are expanding their enrollment plans, particularly focusing on urgently needed majors such as artificial intelligence and renewable energy [12][15] - Notable universities like Tsinghua University and Peking University have announced increases in enrollment numbers, with some institutions expanding by hundreds of students [12] - The Ministry of Education has established a mechanism for the rapid setting of strategic urgent majors, allowing for immediate responses to industry needs [15] Group 3 - The article discusses the expansion of vocational undergraduate programs, with 32 new schools proposed, including 23 vocational universities upgraded from higher vocational colleges [21] - Vocational undergraduate programs are designed to meet the demand for high-level technical talent, closely aligning with local industry needs [21] - The training model for vocational undergraduates emphasizes practical experience through school-enterprise cooperation, with over 50% of study time dedicated to practical training [25]

Core Viewpoint - The article discusses the introduction of dLLM-Cache, a novel inference caching mechanism for diffusion-based Large Language Models (dLLMs), which significantly enhances inference speed without compromising output quality [2][3][21]. Research Motivation - Diffusion-based language models are emerging as a significant paradigm in language generation, showcasing superior capabilities in tasks like "reversing the curse" and mathematical reasoning compared to autoregressive models (ARMs) [8][10]. - However, the inference process of dLLMs typically requires hundreds of denoising steps, leading to substantial computational costs and inefficiencies, particularly since existing acceleration methods like KV Cache are incompatible with the bidirectional attention architecture of dLLMs [10][11]. Method Overview - The authors identified that features of prompt tokens remain stable during the denoising process, allowing for their reuse, while only a small portion of response tokens exhibit significant changes [13][14]. - The V-verify mechanism was introduced to efficiently identify which response tokens require updates based on the changes in their underlying features, achieving a reduction of up to 75% in redundant computations [16][17][20]. Experimental Results - The effectiveness of dLLM-Cache was rigorously tested on LLaDA 8B and Dream 7B models across various benchmark tasks, demonstrating over 5 times acceleration in inference speed while maintaining or slightly improving model performance [21][25]. - In specific tasks like HotpotQA, dLLM-Cache achieved a remarkable 9.1 times speedup without loss of quality, showcasing its robust performance across different dLLM architectures [21][28]. General Applicability - The dLLM-Cache method was successfully applied to different dLLM architectures, confirming its versatility and effectiveness in enhancing inference efficiency across various models [25][28].

导读：近日，上海交通大学发布多批政府采购意向，仪器信息网特对其中的仪器设备品目进行梳理，统计出33项仪器设备采购意向， 预算总额达1.04亿元。 特别提示 微信公众号机制调整，请点击顶部"仪器信息网" → 右上方"…" → 设为 ★ 星标，否则很可能无法看到我们的推送。 近日，上海交通大学发布33项仪器设备采购意向，预算总额达1.04亿元，涉及自动化蛋白纯化检测系统、扫描电镜制 样系统、高效液相色谱-场流分析仪、高温凝胶渗透色谱仪、多模态荧光红外共聚焦显微成像系统等，预计采购时间为 2025年3~6月。 | 采购 | | 采 预算 购 | | --- | --- | --- | | | 需求概况 | | | 项目 | | 万元 时 | | | | 间 | | 自动 | | 202 | | 化蛋 | | | | 白纯 | 设备，1套。包括1. 高速冷冻离心机；2. 低速冷冻离心机；3. 定制超声破碎仪；4 过柱模块；5 蠕动泵；6 超滤收集模块；7 轨道式六 | 5 | | | 轴机械臂；8 中控系统；9 制冷样本架；10 自动化平台等模块，能够实现蛋白质样品的自动化纯化与检测。具体事宜由成交供应商按采 | 3 ...

避免高校"唯AI率"，要强化数据治理，通过纳入多元原创语料、提升检测模型适配能力等方式，让AI查 重"聪明"起来，防止出现"原创作品被检出高AI率、AI生成内容绕过查重"的现象。更重要的是，要合理 评价学术贡献，应适当弱化对AI率的权重。毕竟，论文的价值在其"原创性贡献"，AI率不能定其学术价 值的高下。 今年年初，上海交大就发布AI的使用规范，将高等教育教学领域的人工智能应用划分为禁止使用、有 限使用、鼓励使用、开放使用四种类型。对风险较低或师生经讨论达成较高共识，确保使用符合教育教 学目标、伦理规范和技术可靠性要求的AI工具，学校甚至鼓励学生使用。这种探索值得借鉴。培养学 生的AI素养，循序渐进明确AI在学术写作中的使用规范与边界，构建更为科学、包容的论文评价体 系，或许比要求学生一味降低AI率更为妥帖。 毕业季到来，不少学校要求论文进行AI查重。不过，有高校教师吐槽，明明是学生自己的原创论文， 却被检测出"AI率过高"，为通关不得不重新修改以降低"AI率"，"弄得学生无法正常表达了"。甚至，也 有学生表示，面对误判，自己只能通过加入冗余表达、牺牲表达完整性等方式，来获得"理想的AI率"。 AI查重，无法 ...

5月22日，广州日报数字化研究院（GDI智库）发布"GDI大学一流学科排行榜（2025）"，对全国534所 院校的92个学科（人文社会学科20个、自然科学学科34个、工程学科38个）进行科学评价、排行，范围 涵盖上述学科设有硕士点的全部院校（不含香港特别行政区、澳门特别行政区和台湾地区院校，未涉及 军事类院校）。 GDI大学一流学科排行榜（2025）构建"人才培养""科学研究""学科声誉""二次评估"四维评价体系，坚 持定量评价和定性评价相结合，对相关院校学科发展进行独立、科学评价。榜单坚持分类评价，对人文 社会学科、自然科学学科和工程学科采用不同评价标准，以求更好体现学科特色、反映学科内涵特征， 科学精准地对学科进行差异化评价。 一般认为，我国一所大学的某个学科排名进入全国排名前5，表明该大学这一学科的发展水平已跨入"世 界一流学科"行列或接近"世界一流学科"水平；进入全国排名前25表明已进入"国内一流学科"行列；进 入全国排名前50表明已进入"国内高水平学科"行列。 据此分析统计，"世界一流学科"分布在全国134所院校，其中103所为"双一流"建设高校。在拥有"世界 一流学科"数量上：北京大学有39个，排 ...

Group 1 - The ASC World University Supercomputing Competition (ASC25) concluded on May 14 in Qinghai, with Shanghai Jiao Tong University and Peking University winning the first and second places respectively [1][2] - The competition, initiated by China, is the largest university supercomputing contest globally, attracting over 300 teams from around the world to participate [1][4] - The event lasted nearly five months, with the final competition held at Qinghai University from May 10 to 14, featuring 25 teams that advanced from the preliminary rounds [1] Group 2 - Shanghai Jiao Tong University excelled in multiple topics such as AlphaFold3 inference optimization and cosmic neutrino detection simulation, securing the championship [2] - Peking University performed well in RNA methylation modification site detection, earning the runner-up position [2] - A super team composed of Sun Yat-sen University, Peking University, Beijing Normal University, Fuzhou University, and National Taiwan University received the Super Team Award [2] Group 3 - The president of Qinghai University emphasized the importance of integrating supercomputing with artificial intelligence for talent development, highlighting the need for collaborative innovation in algorithms and computing power [4] - The ASC competition aims to promote the cultivation of young talents in supercomputing through technological competitions and cutting-edge applications, having held 12 editions to date and attracting thousands of university students from six continents [4]

为了吸引人才，上海杨浦区放大招，毕业生"短期免租"、优秀UP主购房最高获200万补贴…… 5月8日，2025年杨浦区人才高质量发展大会举行。会上推出《"优享计划"的实施细则》《"短期免租"试行方案》《区校人才发 展服务联盟章程》等多项人才举措，提升城市"青和力"，助力杨"数"浦发展。 创作者"优享计划" 会上，杨浦区发布《杨浦区关于全球互联网优质内容创作者"优享计划"的实施细则》。以真金白银支持优质内容创作蓬勃发 展，推动有影响力和代表性的UP主等创作人才成为青年群体中主旋律的传播者和优秀文化的弘扬者。 据介绍，《实施细则》推出孵化平台共建、创作激励支持、人才薪酬奖励等十条举措。安居方面，对创作者在本区内购买首套 房给予一次性50万元到200万元住房补贴。入住高品质人才社区优先供给房源，给予每月最高8000元租房补贴。薪酬奖励方 面，未纳入个人所得税税负差额补贴及本市重点产业领域专项奖励政策的创作者，年度薪酬达到100万元以上，给予其薪酬收 入5%到8%的奖励。 落户居留便利方面，国内创作者优先给予落户资格，其配偶、未成年子女按规定随迁随调。外籍创作者在通关通行、在华工 作、证照办理等方面给予便利。综合保障方 ...

中新网上海5月9日电 (记者 许婧)上海交通大学9日宣布，作为上海交大"AI+"战略布局的延展和在国际 化方向的先行试点，上海交大巴黎卓越工程师学院携手法国两大顶尖高校集团——巴黎综合理工大学与 巴黎文理研究大学，推出面向2025级本科生的首届"笛卡尔班"。 该班级以"数理+AI+工程"三位一体的人才培养体系，厚植数理基础、聚焦AI底层逻辑、强化AI素养与 全球治理能力，探索人工智能时代拔尖创新人才培养的新范式，致力于培养面向世界和未来的新时代全 球化AI拔尖人才。 人工智能正在深刻重塑全球科技与产业格局，其核心在于基础理论、底层算法与工程实现的深度融合。 法国在数理教育领域长期处于世界领先地位，拥有严密的知识体系、深厚的历史沉淀和卓越的人才培养 传统，法国数学家连续20年以上获得菲尔兹奖。与此同时，上海交大在工科领域积淀深厚，注重工程实 践与跨学科融合。 此番推出的"笛卡尔班"以法国数学家、物理学家、哲学家勒内·笛卡尔命名，象征理性思维与跨学科探 索精神，体现项目对AI发展基础性、哲学性问题的关注。 该班级突出数理与AI的底层融合，既继承法式教育在抽象建模与逻辑推演方面的训练优势，又融汇上 海交大工科教育在 ...

Group 1 - General Secretary Xi Jinping affirmed Shanghai's effective development of artificial intelligence (AI) during his visit, emphasizing the rapid iteration and explosive growth of AI technology, and encouraging Shanghai to leverage successful experiences in AI ecosystem development [1] - Shanghai's higher education institutions are increasingly confident and proactive in AI development, with numerous AI research institutions emerging and interdisciplinary AI programs gaining traction [1] Group 2 - The Shanghai Municipal Education Commission highlighted the need for deeper talent cultivation reforms and research paradigm changes to keep pace with the rapid technological revolution led by AI, advocating for tailored reforms in universities to align with economic and industrial needs [2] - Shanghai Jiao Tong University aims to align its academic focus with national strategic needs, utilizing AI as a key to overcoming critical challenges and fostering interdisciplinary integration [2] - East China Normal University plans to enhance its academic offerings in response to national strategic demands, focusing on emerging fields like quantum science and technology, while promoting AI integration across all disciplines [2] Group 3 - Shanghai University of Technology is committed to implementing AI in educational reforms, aiming to cultivate top engineering talents through a comprehensive approach that includes curriculum design and international collaboration [3] - Tongji University is establishing several research institutes focused on future disciplines, driven by AI and technology innovation, to enhance its status as a world-class university [4] - Fudan University is strategically developing new engineering innovation colleges to foster original innovation capabilities in technology talent, aligning with national strategic needs [4] Group 4 - Shanghai Second Polytechnic University recently established an AI research institute, marking a significant step in its AI research and development efforts, with a comprehensive curriculum that integrates AI across all undergraduate programs [5] - AI is recognized as a crucial engine for upgrading traditional engineering technologies, with applications in complex environments such as shield tunneling, enhancing risk management and automation [7] - Shanghai is actively promoting AI's role in reshaping higher education and is designated as a pilot area for national AI talent cultivation reforms, fostering an environment for innovative research practices [7]