Core Insights - The research highlights a new threat to autonomous systems, specifically "visual attacks" that can hijack decision-making processes of self-driving cars and drones through malicious text embedded in the environment [1][2][3] - The study emphasizes the urgent need for the industry to establish new safety standards and protective mechanisms to counteract these vulnerabilities [1][2] Group 1: Research Findings - Researchers from the University of California, Santa Cruz, have demonstrated that attackers can manipulate autonomous systems by embedding specific text in physical objects like road signs and posters, leading to dangerous behaviors [2][3] - A framework named "CHAI" was developed to test this concept, which optimizes attack text using generative AI and adjusts visual attributes such as color, size, and position to enhance the effectiveness of the attack [2] - In tests, the CHAI attack successfully interfered with navigation judgments of self-driving vehicles, achieving a manipulation success rate of up to 95.5% in simulated drone scenarios [2] Group 2: Implications for the Industry - The findings indicate that such attacks are feasible in the physical world, posing a real threat to the safety of intelligent systems as AI becomes more integrated into physical environments [2] - The research serves as a warning for the industry to consider the broader implications of AI safety and to conduct more proactive studies to strengthen the security foundation of these technologies [3]

论文第一作者、北京通用人工智能研究院张驰博士介绍，TongGeometry能从浩如烟海的空间组合中， 精准捕捉到具备人类数学家审美标准的高质量题目，在国际上首次实现从"模仿解题"到"自主创造"的范 式转变。 我国科研团队近日开发出全球首个同时具备自主出题和自动解题双重能力的通用人工智能系统——"通 矩模型"（TongGeometry）。相关成果"基于引导树搜索的奥数几何问题提出与解答系统"1月26日发表于 《自然·机器智能》上。 奥林匹克数学竞赛被视为人工智能逻辑推理能力的"试金石"。2024年初，DeepMind公司开发的 AlphaGeometry人工智能系统展示了AI在解题方面的巨大潜力，但其本质上是一个"被动解题者"，训练 极度依赖大规模的合成数据和昂贵的计算资源。与之相比，我国自研的TongGeometry则展现出更高维 度的智能：不仅是一个能满分交卷的"优等生"，更是一位能创造优美、新颖题目的"出题名师"。其自主 生成的3道几何新题，已正式入选2024年全国中学生数学联赛（北京赛区）及美国精英奥赛。 论文共同通讯作者、北京大学心理与认知科学学院助理教授朱毅鑫表示，这意味着中国科研团队在自动 化推理 ...

Core Insights - The research highlights a new threat to autonomous systems, specifically "visual attacks" that can hijack decision-making processes of self-driving cars and drones through malicious text embedded in the environment [1][2] Group 1: Research Findings - Scientists from the University of California, Santa Cruz, have revealed that attackers can manipulate autonomous systems by embedding specific text information in physical environments, leading to dangerous behaviors [1] - The study introduces the concept of "environmental indirect prompts," where malicious text can be placed on road signs or posters to mislead AI systems that rely on visual language models [2] - A framework named "CHAI" was designed to demonstrate command hijacking of embodied AI, optimizing attack text using generative AI and adjusting visual attributes to enhance attack effectiveness [2] Group 2: Experimental Results - The CHAI attack framework was tested in three scenarios: autonomous driving, emergency landings of drones, and target searches, showing a success rate of up to 95.5% in manipulating autonomous systems [2] - In real-world tests, misleading images successfully interfered with the navigation judgments of test vehicles, confirming the feasibility of such attacks in physical environments [2] Group 3: Industry Implications - The findings serve as a warning for the industry, emphasizing the need for new safety standards and protective mechanisms as AI becomes increasingly integrated into physical systems [1][2][3] - The research calls for more comprehensive considerations and proactive studies to ensure the safety of AI technologies as they are deployed in real-world applications [3]

我国科研团队近日开发出全球首个同时具备自主出题和自动解题双重能力的通用人工智能系统 ——"通矩模型"（TongGeometry）。相关成果"基于引导树搜索的奥数几何问题提出与解答系统"1月26 日发表于《自然·机器智能》上。 相比AlphaGeometry需要庞大的算力集群，TongGeometry仅需单张消费级显卡即可在最多38分钟 内，解决近25年所有的奥数几何难题。 论文共同通讯作者、北京大学心理与认知科学学院助理教授朱毅鑫表示，这意味着中国科研团队在 自动化推理的逻辑核心领域实现关键技术自研，并在性能与功能多样性上全面超越以DeepMind为代表 的国际顶尖水平。同时，我们的系统在理解逻辑底层美学和自主发现科学规律方面走在了前列。这种不 依赖海量标注数据、通过内部逻辑自我演化的路径，正是通用人工智能（AGI）发展的关键。 奥林匹克数学竞赛被视为人工智能逻辑推理能力的"试金石"。2024年初，DeepMind公司开发的 AlphaGeometry人工智能系统展示了AI在解题方面的巨大潜力，但其本质上是一个"被动解题者"，训练 极度依赖大规模的合成数据和昂贵的计算资源。与之相比，我国自研的TongGeom ...

Group 1 - The core viewpoint of the articles highlights the rapid integration of AI, particularly large language models (LLMs), into business operations, which brings both transformative potential and significant security risks [1] - AI browsers, such as OpenAI's ChatGPT Atlas and Perplexity's Comet, are set to revolutionize user interactions by automating tasks like form filling and booking, but they also introduce new vulnerabilities that could lead to data breaches and unauthorized actions [2] - Security experts emphasize the need for proactive measures in AI governance, including unique identification for AI agents, data classification, and emergency shutdown mechanisms to mitigate risks associated with AI's increasing autonomy [3] Group 2 - Prompt injection attacks, which manipulate LLMs to bypass security protocols and leak sensitive information, have been identified as a top threat by organizations like OWASP, highlighting the need for robust defenses against such vulnerabilities [4] - The evolution of security access service edge (SASE) into AI-aware access architecture is crucial for managing AI traffic and ensuring compliance, marking a shift from passive to active defense strategies in AI security [5][6] - The establishment of AI security posture management (AI-SPM) systems is anticipated to provide centralized monitoring and governance of AI models and data, ensuring compliance with international risk management frameworks and enhancing overall security [6]

制造出"永不沉没"的船只是工程师们一直以来的希望。为实现这一目标，美国罗切斯特大学光学研究所 研究人员开发出一种新技术，可以将普通的金属管变成永不沉没的物体。这意味着无论其在水中浸泡多 长时间，或遭受多大损坏，都能保持漂浮状态。 研究人员通过蚀刻铝管的内壁，在其表面创造出微米和纳米级的坑洞，使其表面具备超疏水性，能够有 效排斥水分，保持干燥。当经过处理的管子进入水中时，超疏水表面会在管内形成稳定的气泡，从而防 止管道被水浸泡并沉入水中。这种机制与潜水钟蜘蛛和火蚁在水中形成浮筏保持浮力的原理类似。 研究人员还在管子中间加入了隔板，即使将管子垂直推进水中，气泡也能保持在管内，从而保持其浮 力。 此前，研究人员已在2019年首次展示了类似的超疏水浮力装置，使用了两块超疏水圆盘，将其密封在一 起，利用气泡产生浮力。但该设计在极端角度下容易失去浮力。相比之下，目前的管道设计在多种复杂 环境下表现更为稳定，尤其在海洋湍流等条件下，管子能持续保持漂浮。 研究人员在恶劣的测试环境中对管子进行了数周的连续测试，发现即便是管道遭受大范围破损，依然能 浮在水面。 此外，多个超疏水管道可联结成浮筏，用于船只、浮标和浮动平台的构建。使 ...

Core Insights - The innovative drug Shuwotini tablets, which targets a rare mutation in non-small cell lung cancer, has won the first prize in Jiangsu Province's Science and Technology Progress Award, highlighting its significant breakthrough in original innovation and clinical application [1] - Shuwotini tablets are the first and only drug approved for the EGFR 20 insertion mutation, receiving market approval in China in August 2023 and subsequently being included in the national medical insurance directory, significantly reducing patient costs [1] - The award recognizes the project's contributions to technological innovation and its impact on economic and social development, emphasizing the comprehensive strength of the research and development company in key technology breakthroughs and results transformation [1] Company and Industry Developments - The company, Dizhe Pharmaceutical, aims to continue addressing unmet clinical needs through technological innovation, accelerating the delivery of globally competitive original innovations to benefit patients worldwide [2] - The drug is expected to gain international recognition, as it was approved by the U.S. FDA in 2025, marking it as the first independently developed innovative drug from China to be approved in the U.S. and included in the National Comprehensive Cancer Network guidelines for non-small cell lung cancer [1]

Core Insights - The research team led by Hu Haike at the Hangzhou Institute of Medicine has developed the world's first diagnostic kit for distinguishing between benign and malignant lung nodules, which has received regulatory approval as a Class III medical device in China [1][2] - Lung cancer remains the most prevalent and deadly cancer in China, with approximately 150 million people suffering from lung nodules, of which over 95% are benign [1] Group 1: Product Development - The team has focused on early lung cancer diagnosis since 2016, utilizing tumor-specific antibody detection technology to identify specific biomarkers [2] - They have screened over 400 key proteins related to early lung cancer using high-throughput screening techniques and artificial intelligence, resulting in a combination of 13 optimal diagnostic markers, including 8 novel markers [2] Group 2: Clinical Testing and Performance - The diagnostic kit was tested in multiple hospitals, involving 1,463 patients with lung nodules, of which 794 were lung cancer cases, with early-stage lung cancer samples accounting for 58.19% [2] - The kit demonstrated a sensitivity of over 65% for early lung cancer detection, significantly outperforming traditional tumor markers, and can increase diagnostic accuracy for atypical small nodules to over 85% when used alongside low-dose spiral CT [2] Group 3: Market Potential - The diagnostic kit is expected to be launched in the market after the Spring Festival, providing crucial support for China's "Healthy China 2030" cancer early screening strategy [2]

变压器作为特高压、高压输电系统的核心组件，是保障电网安全稳定的关键。在高电压电网中，其特殊 的绕组结构易由轻微匝间故障快速发展为严重故障，进而引发爆燃起火，严重威胁主设备和系统安全运 行。现有技术受到故障过程多次复杂变化和计算算法影响，动作时间长达50—60毫秒，急需大幅提升动 作性能，降低设备高电压等级下的故障能量累积。 记者1月26日从中国电力科学研究院获悉，近日，该院自主研发的"高性能变压器匝间保护技术"实现转 化，并在国家电网公司特高压、高压直流工程换流站大规模应用，成功攻克困扰行业多年的技术难题， 为我国特大型交直流电网安全运行筑牢屏障。该项成果获得国家电网公司2025年科学技术进步奖一等 奖。 该项目负责人表示，该成果实现了从理论研究、技术突破、样机研制到规模化应用的全链条创新，推动 了国产高端装备保护技术升级，显著提升了变压器安全运行水平。在"双碳"目标引领下，该项核心技术 与装备将持续守护万家灯火，为我国特大型交直流电网安全稳定运行提供坚实支撑。 责任编辑：王奕博 为破解这一痛点，在国家电网公司总部指导下，中国电力科学研究院组建专项攻关团队，聚焦变压器保 护原理等关键难点开展攻坚。在多年技术研 ...

Core Viewpoint - The recent outbreak of Nipah virus in West Bengal, India, has raised international concern due to its high fatality rate and human-to-human transmission potential, prompting preventive measures in neighboring countries and within China [1] Group 1: Virus Overview - Nipah virus is not a new virus; it was first reported in Malaysia in 1988 and has since been detected in several countries including Bangladesh, Cambodia, and India [2] - India has experienced Nipah virus outbreaks before, with four confirmed cases reported in May-July 2022, resulting in two deaths [2] Group 2: Transmission and Prevention - The initial case in India was linked to the consumption of fresh date palm juice contaminated by bats, which are the natural hosts of the virus [3] - Transmission occurs primarily through animal-to-human and human-to-human contact, with close contact posing a higher risk for healthcare workers and family members of infected individuals [3] - Public health recommendations include avoiding raw fruit juices, maintaining distance from bat habitats, and ensuring proper hygiene when consuming fruits [4] Group 3: Monitoring and Detection - Currently, there are no reported Nipah virus cases in China, but the virus has been included in the country's monitoring system [5] - Hospitals in China are equipped with next-generation sequencing technology to detect various pathogens, including Nipah virus, and a three-tiered detection system is in place for suspected cases [5] - Specialized testing kits for Nipah virus have been developed and are available for screening at ports of entry and for clinical diagnosis [5]