马志宇表示，一方面，从数据维度跳出"通用数据池"，补充细分场景信息，除了净利润、ROE在内的公开财报核心指标等传统基本面因子外，灵均投资会主 动拓展更细分的数据来源，例如与券商研究所合作开展"阿尔法捕获"（Alpha Capture），直接引入分析师对行业、公司的点状研究信息，将这些非标准化的 洞察转化为可量化的信号。另一方面，在方法论维度，灵均投资根据行业特性做精细化设计，同时结合机器学习，突破传统因子的局限性。 10月15日下午，由中国证券报主办、国信证券独家冠名的"固本 砺新 行远——2025私募基金高质量发展大会"在深圳举行。在"AI引领变革 量化投资的崛起 与未来"圆桌论坛上，灵均投资首席投资官马志宇表示，面对量化行业的同质化挑战，灵均投资通过长期深耕，在数据端拓展细分信息，在方法端实现"行业 定制+特征化建模"，既避免了和同行在通用因子上的同质化竞争，又能让基本面因子在不同场景下更精准地发挥作用。 灵均投资首席投资官马志宇在圆桌论坛环节发言 ...

Core Viewpoint - The article emphasizes the transformative impact of quantitative research on the asset management industry amidst a rapidly changing global macroeconomic landscape and increasing volatility in international financial markets [1]. Group 1: Event Overview - The event will feature discussions on macro quantitative scenario analysis, risk budgeting applications in the Chinese market, and Bloomberg's portfolio management and factor model solutions [1]. - A roundtable forum will explore how experiences from mature overseas markets can empower the development of quantitative strategy indices in China [1]. Group 2: Key Speakers - Notable speakers include Li Yongjin from CITIC Securities, Arun Verma from Bloomberg, Sue Li from Bloomberg, Wayne Curry from Bloomberg, and several other experts from Bloomberg's global and China teams [2][6]. Group 3: Topics of Discussion - The agenda includes leveraging machine learning strategies based on macro drivers and the application of cutting-edge intelligent technologies in quantitative research [1]. - The event aims to provide insights into best practices for asset allocation and risk management tailored to the characteristics of the Chinese market [1].

IBIW INTERNATIONAL BIOPHARMA INDUSTRY WEEK SHANGHAI 2025 阜 年 監管財能 政策动向与全球协同 Atonio Barra Torr an All Produktion Comment Co 66 祝贺上海成功打造了这么一个 重要的平台,在"链动全球 赋能产业" 的主题下,产业周链接了科学、产业、 金融、政策。 B 蓝恭涛 国家药监局药品注册管理司 副司长 66 总体目前支持医药研发创新的 政策制度体系日趋完善,以"药品注 册管理办法"为核心形成的规范性文 件60多个,形成技术指导原则500多 个,鼓励创新的氛围日渐形成。 2025上海国际 International Biopharma Indu 2025 上海国际生物医药 产业周 66 有很多英国和伦敦的机构在上 海生命医药领域都有投资,他们很喜 欢上海的生物医疗生态体系。生命医 药领域潜力无穷,我们也很期待和上 海企业加强合作。 S 新集馆 原始创新与技术突破 sia edCity) 播 enemzs 士理学成医学 66 我们非常乐观,这些针对HIF-1 α、HIF-2α的药品将会对治疗癌症病 人很有 ...

Core Viewpoint - The company has over 20 years of experience in the bulk commodity information service sector, accumulating a vast amount of price and fundamental data [1] Group 1 - The data and information content is collected and written by a professional analyst team, ensuring authority, timeliness, and accuracy [1] - The company has established a data center within its software industrial park, capable of managing the entire data lifecycle from collection to application [1] - The company employs machine learning and deep learning algorithms to train, evaluate, optimize, and persist data sets, supporting business user modeling and forecasting needs [1]

Core Insights - Astera Labs has seen its stock price surge by 250% over the past six months, making its co-founders billionaires amid the AI infrastructure boom [1] - The company focuses on connectivity technology to enhance AI infrastructure, with significant revenue growth projected [2] - Astera Labs' revenue has increased over 11 times from $35 million in 2021 to an expected $396 million in 2024, with profitability anticipated in 2025 [2] Company Background - The co-founders, Jitendra Mohan and Sanjay Gajendra, previously worked at Texas Instruments and National Semiconductor before founding Astera Labs in 2017 [3] - They identified a gap in connectivity technology that was not keeping pace with advancements in AI and machine learning [3] - The company raised $50 million in 2021 at a valuation of $950 million and later secured $150 million in 2022, increasing its valuation to $3.2 billion [3] Financial Performance - Astera Labs went public in March 2024, raising $820 million and achieving a market capitalization of approximately $6 billion [3] - The current market valuation has soared to $34 billion, with the co-founders holding about 4% of the shares, valued at $1.5 billion each [4] - The co-founders have sold over $200 million worth of stock since the IPO [4]

Core Viewpoint - Bond investment is shifting from "trend dividends" to "structural dividends" due to the resonance of the current economic cycle, policy tools, and market sentiment [1] Recent Interest Rate Trends - Economic cycle determines the upper limit of interest rates, with China's manufacturing PMI at 49.5% and PPI declining by 3.3% year-on-year, indicating weak internal and external demand [2] - Monetary policy is reshaping the formation mechanism of interest rate centers, with a total of 9 reserve requirement ratio cuts and a cumulative reduction of 80 basis points in the 7-day reverse repurchase rate since January 2021 [3] - Market sentiment is lowering the lower limit of interest rates, with significant fluctuations driven by investor sentiment and geopolitical factors, leading to short-term interest rate volatility [4] Future Investment Opportunities - The rise of quantitative trading is transforming bond trading from automation to intelligence, with a focus on data-driven trading models [5][6] - The layout of green and technology innovation assets is becoming crucial, with green bond issuance reaching 2.5 trillion yuan and technology innovation bonds seeing an average cost of 3.2% [7] - The cross-border bond market is expanding, with the issuance of dim sum bonds in Hong Kong reaching 1.2 trillion yuan, a 27% increase year-on-year, and a significant return on Chinese dollar bonds [8] Market Outlook and Innovative Development - The offshore bond market is expected to expand, enhancing Shanghai's role as an international financial center and facilitating cross-border financing [9][10] - The REITs market is rapidly developing, with a total market value exceeding 200 billion yuan and annualized dividend yields stable at 3%-5% [11][12] - The variety of floating rate bonds is increasing, with a current stock of approximately 605.4 billion yuan, representing only 0.38% of the bond market [13] Conclusion - The bond market is in a complex phase of multi-factor resonance, with macroeconomic expectations, policy goals, and market sentiment contributing to underlying market volatility [14]

Core Viewpoint - The article discusses Geoffrey Hinton's perspective on artificial intelligence (AI), suggesting that AI may already possess a form of "subjective experience" or consciousness, albeit unrecognized by itself [1][56]. Group 1: AI Consciousness and Understanding - Hinton posits that AI might have a nascent form of consciousness, which is misunderstood by humans [2][3]. - He emphasizes that AI has evolved from keyword-based search systems to tools that can understand human intentions [10][14]. - Modern large language models (LLMs) exhibit capabilities that are close to human expertise in various subjects [15]. Group 2: Neural Networks and Learning Mechanisms - Hinton explains the distinction between machine learning and neural networks, with the latter inspired by the human brain's functioning [17][21]. - He describes how neural networks learn by adjusting the strength of connections between neurons, similar to how the brain operates [21][20]. - The breakthrough of backpropagation in 1986 allowed for efficient training of neural networks, significantly enhancing their capabilities [38][40]. Group 3: Language Models and Cognitive Processes - Hinton elaborates on how LLMs process language, drawing parallels to human cognitive processes [46][47]. - He asserts that LLMs do not merely memorize but engage in a predictive process that resembles human thought [48][49]. - The training of LLMs involves a cycle of prediction and correction, enabling them to learn semantic understanding [49][55]. Group 4: AI Risks and Ethical Considerations - Hinton highlights potential risks associated with AI, including misuse for generating false information and societal instability [68][70]. - He stresses the importance of regulatory measures to mitigate these risks and ensure AI aligns with human interests [72][75]. - Hinton warns that the most significant threat from advanced AI may not be rebellion but rather its ability to persuade humans [66]. Group 5: Global AI Landscape and Competition - Hinton comments on the AI competition between the U.S. and China, noting that while the U.S. currently leads, its advantage is diminishing due to reduced funding for foundational research [78][80]. - He acknowledges China's proactive approach in fostering AI startups, which may lead to significant advancements in the field [82].

Core Viewpoint - Andrew Tulloch, co-founder and chief architect of Thinking Machines Lab, has left the company to join Meta, despite previously rejecting a $1.5 billion compensation package from Meta [1][18]. Group 1: Andrew Tulloch's Background and Career - Tulloch has a strong academic background, graduating with honors in mathematics and statistics from the University of Sydney and later earning a master's degree in mathematical statistics and machine learning from Cambridge University [8][11]. - He began his career at Goldman Sachs, developing financial products and trading strategies before moving to Facebook (now Meta) in 2012, where he worked for 11 years in machine learning [10][11][6]. - Tulloch's expertise in machine learning was further utilized at OpenAI, where he worked on training models like GPT-4.5 before co-founding Thinking Machines Lab [16][15]. Group 2: Transition to Meta - Tulloch's return to Meta is seen as a "homecoming," as he had previously spent a significant amount of time there [6]. - His departure from Thinking Machines Lab was described as a personal decision, and there was speculation about the reasons behind it, especially given the company's high valuation of $12 billion [4][21]. - The recruitment efforts by Meta included a direct approach from CEO Mark Zuckerberg, who initially sought to acquire Thinking Machines Lab before focusing on hiring Tulloch and other employees [19][20]. Group 3: Compensation and Market Dynamics - Tulloch had previously turned down a $1.5 billion offer from Meta, which included stock options, indicating a potential increase in compensation that may have influenced his decision to join [18][19]. - The article hints at the possibility that Tulloch's compensation package may have increased to $2 billion, reflecting the competitive nature of talent acquisition in the tech industry [21].

Core Insights - The collaboration between AMD and Sony Interactive Entertainment (SIE) has initiated the "Project Amethyst," aiming to integrate AI and machine learning into future gaming hardware and software [1][2] - The partnership will develop foundational AI systems for PC and PlayStation platforms, symbolizing a deep technological integration between AMD's red and PlayStation's blue, resulting in a purple crystal color [1] - The AI advancements will enhance not only image reconstruction but also explore neural frame generation and ray regeneration to improve real-time ray tracing and path tracing efficiency [2] AMD's Next-Generation GPU Technologies - AMD has revealed three core technologies for its next-generation GPU architecture, focusing on AI acceleration, ray tracing efficiency, and memory optimization [3] - The neural array technology is a key strategy for AMD to compete with NVIDIA's Tensor Core, indicating that AI is becoming a core part of the architecture [3] - The "Project Amethyst" will allow AMD and Sony to validate and promote these technologies on millions of gaming consoles, enhancing AMD's competitive edge in ray tracing performance and AI-driven features [3][4] Hardware Upgrades and Innovations - Sony has disclosed hardware upgrades for its next-generation console, with a focus on the collaboration project "Project Amethyst" [4] - Key innovations include Radiance Cores for efficient ray tracing and path tracing, neural arrays for AI rendering, and universal compression technology to reduce memory bandwidth requirements [4][6] - These advancements suggest that the next Sony gaming console, likely the PS6, will be capable of running ray tracing and path tracing games comparable to current PC graphics hardware [6][7] Future Implications - The technologies developed under "Project Amethyst" are expected to be implemented in future AMD GPUs and SoCs, with a timeline indicating these innovations will appear in upcoming gaming consoles in the coming years [7] - The collaboration signifies a strategic alignment between AMD and Sony, enhancing the gaming experience through AI and advanced graphics technologies [5][6]

为应对这一挑战，研究团队提出了一种结合物理模型与机器学习技术的新路径。该方法基于"自动微 分"技术——一种常用于深度学习中计算导数的算法，用于追踪输入变量微小变化对输出的影响。他们 利用这一机制，在分子动力学模拟框架下直接优化氨基酸序列，使其具备预定的物理或功能特性。与依 赖大量数据训练的典型人工智能模型不同，该方法依托已有且足够精确的物理模拟体系，通过梯度优化 高效搜索满足特定功能需求的蛋白质序列，如形成柔性连接结构或响应环境变化的能力。 团队强调，目标并非用数据驱动模型替代物理理解，而是将真实的分子行为规律嵌入设计过程，使生成 的蛋白质序列不仅具备功能性，而且其设计过程本身就根植于自然界真实的动力学原理。由此设计出的 蛋白质是"可微分的"，意味着每一步优化都建立在对系统物理状态连续、精确调控的基础上，而非依赖 黑箱式的预测。 （文章来源：科技日报） 美国哈佛大学与西北大学研究团队合作，开发出一种新型机器学习方法，能够从无序蛋白质中排序，设 计出具有特定性质的内在无序蛋白质（IDPs），从而突破了当前人工智能（AI）工具在解析约30%人类 蛋白质结构上的局限。该成果发表于最新一期《自然·计算科学》。 这类蛋 ...