CoreCognition团队 投稿 量子位 | 公众号 QbitAI 一篇被Yann LeCun转发的ICML 2025研究给了 多模态大模型 当头一棒—— 大部分AI在复杂任务上表现很好，但在人类从小就会的 基础认知能力 上却很拉垮。 研究者建了测评题库 CoreCognition ，覆盖在人类婴幼儿阶段即出现的12种核心认知能力（如客体永恒、视角采择、直觉物理、知觉恒常 等），用来对模型进行系统性测试。 在CoreCognition基准的1503道"经典发展心理学测验"上，230个主流模型系统暴露出对世界常识的"核心知识盲区"。 在归一化准确率对比中，多模态大模型在基础核心认知能力上普遍落后，差距往往达到两位数，即便规模更大也难以弥补。 这是否意味着MLLM（多模态大模型）的先天认知结构中，缺少那些支撑早期人类学习的基础知识机制？ 也就是说，它们是否缺乏"core knowledge"（核心认知能力）？ 构建CoreCognition Benchmark 来自加州大学圣地亚哥分校、约翰霍普金斯大学、埃默里大学、北卡罗来纳大学教堂山分校、斯坦福大学、卡内基梅隆大学等机构的研究人 员，花费一年时间构造并 开 ...

Core Viewpoint - The company, Keda Xunfei, has revised its fundraising plan for the issuance of A-shares, reallocating the use of raised funds to focus on computing power leasing in response to the rapid development of AI and large model technologies [1] Fundraising Plan - The total amount of funds to be raised remains unchanged, with the amount allocated for working capital reduced from 3.2 billion to 800 million [1] - 2.4 billion will be allocated for specialized expenditures related to computing power leasing [1] Computing Power Leasing Project - The company aims to implement a computing power platform project through a leasing model, expanding its computing power scale to support the development of large models and continuous algorithm innovation [1] - This project will enhance the base model upgrades for the Xunfei Spark cognitive large model and provide computing power support for the optimization of various models, including the Spark X1 model and multi-modal models [1] Strategic Implications - The expansion of computing resources is expected to improve the company's capability to provide multi-language large models for global clients, enhancing its technical strength for overseas AI business expansion [1] - The initiative is anticipated to create a new growth point for the company's business revenue [1]

报告在上海人工智能领域筛选出49家先锋成长企业，涵盖娱乐消费、数据、汽车、工业、物流、大模 型、资产管理、教育、金融科技、机器人等细分行业。与生物医药领域类似，浦东新区企业数量仍居首 位，但由于其在张江科学城的集聚程度相对有限，领先优势已明显减弱。 基于"先锋成长企业库"，报告将上海生物医药及人工智能领域表现前十企业评为先锋企业。其中，生物 医药领域先锋企业中，医疗研发外包集群优势明显，动物实验平台成为热门赛道，先进影像设备制造 商、创新药企表现突出；人工智能领域先锋企业正积极布局多模态大模型、无人驾驶底层技术、"未来 产业+AI"融合及垂直应用等多个关键赛道。 报告认为，先锋企业不仅是前沿技术突破的开拓者，更将成为新质生产力的重要载体，推动更多"从0到 1"的原始创新和"从1到100"的产业转化，成为更具开放性、更具国际影响力的创新生态体系中的重要组 成部分。 为此，报告提出"先锋企业"概念。有别于"独角兽"企业、"瞪羚"企业、"猎豹"企业等以财务指标为核心 依据的企业类型，"先锋企业"从企业创新竞争能力视角出发，特指以技术创新为核心驱动力，敢于主动 开辟科技前沿新赛道、发起颠覆性变革的初创科技企业，堪称硬 ...

Core Viewpoint - The company, Torish, reported a significant investment in R&D during the first half of the year, focusing on enhancing its technological foundation and core competitiveness through various initiatives [1] R&D Investment - In the first half of the year, the company's R&D expenditure amounted to 147.51 million yuan, primarily directed towards multi-modal large model optimization, industry-specific large model applications, overseas product development, and strengthening global data collection capabilities [1] Strategic Focus - The company plans to closely monitor market trends and flexibly adjust its business strategies in the second half of the year, with a focus on accelerating the large-scale application of core technologies in benchmark scenarios, deepening industry-specific solutions, improving operational efficiency and profitability, and seizing opportunities presented by data-related policies [1] Business Growth - The company is committed to driving business recovery and growth, aiming to deliver returns to its investors [1]

1 2025年9月26日 ♀ 北京世纪金源大饭店 2025中国安防"人工智能+"行动大会 2025低空安全大会&2025智能机器人论坛 关于大会 人工智能是新一轮科技革命和产业变革的重要驱动力量, 是推动安防行业转型升级的核心引擎。为积极响应国务院常务 会议提出的《关于深入实施"人工智能 +"行动的意见》,贯彻 落实"人工智能 + 安防"行动,以技术革新为引擎,推动安防 新质生产力跃升,兹定于 2025年 9 月 26 日,在北京召开 2025 中国安防"人工智能 +"行动大会。本次大会以"数智世 界 全域安全"为主题,聚焦"人工智能 + 安防"的创新发展, 围绕政策方向、产业变革、技术创新、发展瓶颈等议题展开深 度研讨,旨在凝聚资源、深化交流,探索新赛道新动能,推动 人工智能与安防科技发展深度融合,形成以创新带应用、以应 用促创新的良性循环,支撑和保障国家安全体系与能力现代化, 构建动态敏捷、多元协同的人工智能治理格局。 9 2025年9月26日 2 北京 ·世纪金源大饭店 大会议程 ▼ 金源厅 09:00-12:00 2025 中国安防"人工智能 +"行动大会 09:00-09:05 ● 主持人开场 王 ...

Group 1: Artificial Intelligence Industry Development - The meeting emphasized the need to seize strategic opportunities and implement the "Artificial Intelligence+" initiative, promoting deep integration of AI across various sectors [2] - Focus areas include technology, industry, consumption, livelihood, governance, and global cooperation, aiming to empower numerous industries with AI [2] - The establishment of high-level AI industrial parks and the cultivation of key industries such as smart terminals, networks, and equipment were highlighted [2] Group 2: Quality Improvement Strategy - The meeting stressed the importance of a quality-first mindset and the implementation of a quality enhancement strategy to develop new economic advantages centered on technology, standards, brands, quality, and services [2] - Strengthening quality management and building a robust quality management service platform were key points discussed to enhance the influence and reputation of Wuhan's brands [2] Group 3: Aviation Development - The meeting called for leveraging the city's position as an international comprehensive transportation hub, focusing on market demand and effective management of aviation development funds [3] - The goal is to create a dual hub for international air passenger and cargo transport, promoting high-quality development of the civil aviation industry in the city [3]

Core Viewpoint - The automotive industry is transitioning towards becoming "intelligent entities with warmth," emphasizing the integration of AI technology to enhance user experience and operational efficiency [3][4]. Group 1: AI Application in Automotive Industry - The discussion at the forum highlighted the importance of AI in transforming vehicles into intelligent companions rather than just functional tools [3]. - Dongfeng Motor Corporation has implemented a comprehensive AI strategy, focusing on user scenarios, foundational models, and data-driven approaches to create intelligent automotive experiences [3]. - The company has developed a multimodal large model to capture driver physiological and emotional states, adjusting the in-car environment accordingly [3]. Group 2: Automation in Manufacturing - China has become the largest industrial robot market globally, with the automotive sector being a key area for automation [6]. - The automation rates in stamping, welding, and painting processes exceed 90%, with advancements in humanoid robots changing the landscape of manufacturing [6]. - AI and 3D vision technologies are being integrated into manufacturing processes for quality control and precision tasks [6]. Group 3: Risks and Governance in AI - The application of AI in the automotive sector raises significant safety risks, necessitating a robust governance framework [7][8]. - The risks are categorized into inherent system risks and application-derived risks, including algorithm transparency and data security [7]. - China is enhancing its policy framework for intelligent connected vehicles, focusing on safety management and technical guidelines [8]. Group 4: Future Directions and Consensus - The forum reached a consensus on the need for data-driven collaboration across the automotive value chain to enhance safety and efficiency [10]. - There is a push for deeper integration of AI into core industry processes, moving from isolated applications to systemic integration [10]. - The establishment of a secure and compliant data infrastructure is essential for fostering a collaborative ecosystem in the automotive industry [10].

Group 1 - Sichuan Zigong Transportation Machinery Group Co., Ltd. (运机集团) signed a cooperation framework agreement with iFlytek (科大讯飞) and its subsidiary Zhejiang Tidal Power Technology Co., Ltd. (潮汐力) on September 4 [1] - The strategic cooperation focuses on emerging technologies such as acoustic large models, visual large models, predictive large models, and multi-modal large models, as well as intelligent products like optical fiber voiceprint diagnosis and industrial intelligent robots [1][2] - The collaboration aims to develop and share results in industries including mining, bulk material transportation ports, cement and building materials, metallurgy, electricity, intelligent equipment, and intelligent operations [1] Group 2 - The three parties will work together to seek technology projects, market resources, and business opportunities, establishing a joint development mechanism to address customer needs and industry pain points [1][2] - The partnership emphasizes principles of resource sharing, joint responsibility for results, and win-win cooperation, focusing on AI technology breakthroughs and project development [2] - Future efforts will center on AI large models, AI sensing instruments, and industrial intelligent robots, aiming to create a smarter, safer, and more efficient industrial production system [2]

Core Viewpoint - The company has made significant progress in the field of multimodal large models, applying them across various business lines for enhanced operational capabilities [1] Group 1: Application of Multimodal Large Models - The company utilizes dynamic rules and instructions to implement multimodal large models in behavior recognition, scene analysis, risk warning, and emergency command [1] - Each video is equipped with an intelligent brain through the application of these models, enhancing the understanding of video content [1] - The company has achieved comprehensive perception in video streaming by extracting target event information, creating a complete information cognitive landscape [1] Group 2: Integration with Smart Wearable Devices - The multimodal capabilities have been applied to visual understanding and enhancement in smart wearable devices [1] - The integration of data and service resources has led to a synergy between business scenarios and data capabilities [1]

Group 1 - The core viewpoint of the article highlights that the Hong Kong stock market has shown significant growth this year, with the Hang Seng Index rising over 20% since the beginning of the year, and the Hang Seng Tech Index leading the way [1] - The current valuation of Hong Kong stocks is still considered low compared to historical highs in 2021, indicating potential for further growth, especially in sectors like technology, internet, innovative pharmaceuticals, and new consumption [1][2] - The article emphasizes that external factors, such as the Federal Reserve's potential interest rate cuts, could enhance liquidity and attract foreign capital back to the Hong Kong market, acting as an accelerator for continued market strength [2] Group 2 - Hong Kong's tech companies are becoming increasingly competitive in the AI sector, with recent upgrades to the listing system facilitating the entry of biotech and hard tech companies, thereby enhancing the market's advantages in these areas [2] - There is significant inflow of southbound capital into Hong Kong stocks, with over 900 billion RMB accumulated this year, particularly benefiting large tech stocks like Tencent, Meituan, Xiaomi, and Kuaishou [3] - The article notes a surge in IPO activity in Hong Kong, reflecting the capital market's recognition of its financing capabilities, and suggests that the market may experience a "comeback" this year [3]