在近20年AI发展的路线上，我们正站在⼀个前所未有的转折点。从早期的符号推理到深度学习的突破，再到 如今⼤语⾔模型的惊艳表现， AI 技术的每⼀次⻜跃都在重新定义着⼈类与机器的关系。⽽如今，具身智能正 在全面崛起。 想象⼀下这样的场景：⼀个机器⼈不仅能够理解你的语⾔指令，还能在复杂的现实环境中灵活移动，精确操作 各种物体，甚⾄在⾯对突发情况时做出智能决策。这不再是科幻电影中的幻想，⽽是正在快速成为现实的技术 ⾰命。从Tesla的Optimus⼈形机器⼈到Boston Dynamics的Atlas，从OpenAI的机械⼿到Google的RT-X项⽬，全 球顶尖的科技公司都在竞相布局这⼀颠覆性领域。具身智能的核⼼理念在于让AI系统不仅拥有"⼤脑"，更要拥 有能够感知和改变物理世界的"身体"。这种AI不再局限于虚拟的数字空间，⽽是能够真正理解物理定律、掌握 运动技能、适应复杂环境。它们可以在⼯⼚中进⾏精密装配，在医院⾥协助⼿术操作，在家庭中提供贴⼼服 务，在危险环境中执⾏救援任务。这种技术的潜在影响⼒是⾰命性的：它将彻底改变制造业、服务业、医疗健 康、太空探索等⼏乎所有⾏业。 从顶级会议ICRA 、IROS到Neu ...

Group 1 - The meeting emphasized the importance of high-quality development and the construction of a new development pattern, focusing on the "two highs and four efforts" strategy [1] - The meeting highlighted the need to align with national and provincial "14th Five-Year" plans, conducting thorough research and gathering opinions for informed decision-making [1] - The meeting underscored the significance of capability and style construction as a key approach to achieving high-quality development and effective governance [1] Group 2 - The meeting discussed the promotion of recycled water utilization, advocating for water conservation and sustainable use to support modern urban construction and environmental protection [2] - The meeting addressed strategies to boost consumption, particularly in key areas, enhancing night-time consumption offerings and optimizing service to expand the influence of the "Beautiful Night Zhengzhou" brand [2] Group 3 - The meeting also reviewed preparations for major events such as the 2024-2025 Global Invention Conference National Finals in China [3]

Group 1: Core Perspectives - The construction of a strong education system requires systematic leaps and qualitative changes at both macro and micro levels [1] - Key issues must be addressed by gathering resources such as policies and funding to build a superior ecosystem [1] Group 2: Higher Education Technology Transfer - The Ministry of Education held a national meeting to promote the construction of regional technology transfer centers in universities [3] - Successful integration of technological and industrial innovation relies on strong universities, capable entrepreneurs, and financial support [3] - The first regional center in Jiangsu has demonstrated significant results in gathering various innovation elements and has set a precedent for future centers [3] Group 3: Educational Digitalization - The Ministry of Education emphasized the need for a new learning ecosystem that promotes lifelong learning through collaboration among schools, families, and society [4] - A successful learning society requires sufficient resource supply, convenient learning environments, and comprehensive incentive mechanisms [4] - The educational digitalization strategy is advancing through iterative processes, achieving significant progress in ecosystem integration [4] Group 4: Overall Education Reform - The acceleration of building a strong education system highlights the importance of organized integration and effective mechanism design [5] - There is a need to study non-educational factors that contribute to educational reform and to establish efficient systems and mechanisms [5]

Core Idea - The article discusses the educational philosophy of Grundtvig, a significant Danish thinker who advocated for universal education and lifelong learning, breaking the monopoly of the elite on education and laying the foundation for Denmark's public education system [1][2][3]. Group 1: Grundtvig's Philosophy and Impact - Grundtvig's philosophy emphasizes that education should be accessible to everyone, regardless of their social status, and focuses on awakening the human spirit rather than merely imparting knowledge [2][3]. - He believed that the enlightenment of the populace was essential for national strength and advocated for fundamental reforms in education, establishing adult schools for farmers, workers, and ordinary citizens [2][3]. - Grundtvig founded Denmark's first adult school in 1844, promoting an open educational model where students could choose courses based on their interests, fostering lifelong learning and cultural enrichment [3]. Group 2: Development of Adult Education in Denmark - Under Grundtvig's influence, adult schools proliferated in Denmark, significantly enhancing the cultural quality and economic development of the populace [3]. - The cooperative movement in agriculture, inspired by the education provided in these adult schools, improved product quality and market resilience, contributing to Denmark's modern agricultural success [3]. - Today, Denmark has approximately 70 adult schools, which are integral to the national education system, with a literacy rate of 99% and a higher education enrollment rate of 85% [3].

Core Viewpoint - The article discusses the advancements in end-to-end autonomous driving technology, highlighting the performance of the top competitor, Poutine, in a recent visual-based driving competition, emphasizing its robust training methodology and superior results [1][13]. Group 1: Technical Overview - The leading solution, Poutine, utilizes a 3B parameter Vision-Language Model (VLM) to address long-tail scenarios in visual end-to-end autonomous driving [1]. - The training process consists of two phases: - Phase one involves self-supervised pre-training using a combination of vision, language, and trajectory data, with a total of 83 hours of CoVLA data and 11 hours of Waymo long-tail dataset [2]. - Phase two focuses on fine-tuning through reinforcement learning (RL) using 500 segments of manually annotated data from the Waymo validation set to enhance robustness [2][8]. - The Poutine model achieved a Rater-Feedback Score (RFS) of 7.99 on the Waymo test set, leading the competition [2][13]. Group 2: Data and Methodology - The datasets used include CoVLA, which contains 10,000 front-view images and 30 seconds of driving video, and WOD-E2E, which provides 4,021 long-tail driving scenarios with trajectory information [11]. - The evaluation metric, RFS, is calculated based on the proximity of predicted trajectories to expert-rated trajectories, with a scoring range of 0 to 10 [11]. - The training details include a batch size of 64 and a learning rate of 1e-5 for the CoVLA dataset, while the WOD-E2E dataset used a batch size of 16 with similar training parameters [11]. Group 3: Results and Analysis - Poutine's performance significantly outperformed other models, with a notable score of 7.99, while the second-best model scored 7.91, indicating a substantial lead [13]. - The article notes that while the addition of RL did not drastically improve scores, it effectively addressed challenging scenarios [13]. - The results suggest that the combination of VLM and RL training enhances the model's ability to handle complex driving environments [18]. Group 4: Future Considerations - The article raises questions about the mainstream applicability of VLM and LLM in trajectory prediction, particularly regarding their understanding of the physical world and 3D trajectory information [19]. - It suggests that for conventional evaluation datasets, the advantages of such models may not be as pronounced, indicating a need for further exploration [19]. - The potential integration of action models with VLM for trajectory prediction is proposed as a more comprehensive approach [19].

| 高频选股因子周报（20250623- | [Table_Authors] | 郑雅斌(分析师) | | --- | --- | --- | | 20250627） | | 021-38676666 | | 高频因子上周表现分化，日内收益与尾盘占比因子强势。深度 | 登记编号 | S0880525040105 | | 学习因子依然稳健， AI 增强组合上周表现有所分化。 | | | | | | 余浩淼(分析师) | | 本报告导读： | | 021-38676666 | | 上周（特指 20250623-20250627，下同）高频因子上周表现分化，日内收益与尾盘占 比因子强势。深度学习因子依然稳健，AI 增强组合上周表现有所分化。 | 登记编号 | S0880525040013 | 投资要点： 金融工程/[Table_Date] 2025.06.29 金 融 工 程 周 报 证 券 研 究 报 告 请务必阅读正文之后的免责条款部分 金 融 工 程 [Table_Summary] 高频因子上周表现分化，日内收益与尾盘占比因子强势：日内高频 偏度因子上周、6 月、2025 年多空收益为-0.51%，1.48% ...

以下文章来源于具身智能研究室 ，作者Yuanxq 具身智能研究室 . 分享一些深度强化学习、多/单智能体、具身智能的相关知识。有缘更新，随缘关注。希望大家互相学 习补充。 作者丨 Yuanxq 编辑丨具身智能研究室 点击下方 卡片 ，关注" 具身智能之心 "公众号 >> 点击进入→ 具身 智能之心 技术交流群 更多干货，欢迎加入国内首个具身智能全栈学习社区 ： 具身智能之心知识星球 (戳我) ， 这里包含所有 你想要的。 在复杂动态环境中实现四足机器人的安全高效移动，一直是机器人领域的核心挑战。传统方法 依赖深度相机或中间地图表示，难以应对三维空间中的非平面障碍、空中杂波及动态物体。香 港科技大学团队提出 Omni-Perception 框架，通过直接处理原始 LiDAR 点云数据， 实现了端到端的四足机器人全向避障能力。通过高保真 LiDAR 仿真工具和新型 PD-RiskNet 网络架构，推动了机器人在复杂三维环境中的自主导航技术。 1 、从 LiDAR 点云到全向避障的端到端设计框架 1.Omni-Perception 的核心架构解析 感知-控制一体化设计 优势： 时空信息直接利用 ：避免了点 云到网格 ...

点击下方 卡片 ，关注" 具身智能 之心 "公众号 作者丨 Tian-Yu Xiang等 编辑丨具身智能之心 本文只做学术分享，如有侵权，联系删文 >> 点击进入→ 具身智能之心 技术交流群 更多干货，欢迎加入国内首个具身智能全栈学习社区 ： 具身智能之心知识星球 (戳我) ， 这里包含所有你想要 的。 想象学习走路的情景：尽管祖先的经验让一些与生俱来的能力（例如：平衡感、反应）被编码到我们的 DNA中，但要真正学会走路，仍需要在真实环境中不断练习、摔倒、再爬起。经过一段时间的训练，我们 的大脑和身体会逐渐协调一致，形成与环境交互的策略。这种 由通用能力到特定技能 的转变过程在人类中 十分常见，而如今， 智能机器人 也面临着类似的挑战：即便拥有强大的预训练模型作为"大脑"，在执行具 体复杂任务前，仍需要经过类似于人类学习的"后训练"阶段，才能在新环境、新任务下达到理想表现。 1. 概述 这项工作从 人类运动技能学习 的角度系统性地对总结 VLA模型（视觉-语言-动作模型） 的 后训练（post- training）策略 。其主要贡献如下： (1) 从人类运动学习视角讨论了VLA模型后训练方法 ：将人类运动技能 ...

Core Insights - The article discusses Andrej Karpathy's concept of "Software 3.0," where natural language becomes the new programming interface, and AI models execute specific tasks [1][2]. - It emphasizes the transformative impact of this shift on developers, users, and software design paradigms, indicating a new computational framework is being constructed [2]. Development of LLMs - The evolution of Large Language Models (LLMs) has accelerated since the introduction of the Transformer architecture in 2017, leading to significant advancements in the GPT series and multimodal capabilities [3][5]. - Key foundational papers that established today's AI capabilities are reviewed, highlighting the transition from traditional programming to natural language interaction [5][6]. Foundational Theories - The paper "Attention Is All You Need" (2017) introduced the Transformer architecture, which relies solely on self-attention mechanisms, revolutionizing natural language processing and computer vision [10][11]. - "Language Models are Few-Shot Learners" (2020) demonstrated the capabilities of GPT-3, establishing the "large model + large data" scaling law as a pathway to more general artificial intelligence [13][18]. - "Deep Reinforcement Learning from Human Preferences" (2017) laid the groundwork for reinforcement learning from human feedback (RLHF), crucial for aligning AI outputs with human values [15][18]. Milestone Breakthroughs - The "GPT-4 Technical Report" (2023) details a large-scale, multimodal language model that exhibits human-level performance across various benchmarks, emphasizing the importance of AI safety and alignment [26][27]. - The release of LLaMA models (2023) demonstrated that smaller models trained on extensive datasets could outperform larger models, promoting a new approach to model efficiency [27][30]. Emerging Techniques - The "Chain-of-Thought Prompting" technique enhances reasoning in LLMs by guiding them to articulate their thought processes before arriving at conclusions [32][33]. - "Direct Preference Optimization" (2023) simplifies the alignment process of language models by directly utilizing human preference data, making it a widely adopted method in the industry [34][35]. Important Optimizations - The "PagedAttention" mechanism improves memory management for LLMs, significantly enhancing throughput and reducing memory usage during inference [51][52]. - The "Mistral 7B" model showcases how smaller models can achieve high performance through innovative architecture, influencing the development of efficient AI applications [55][56].

开林企业管理集团有限公司总裁助理郭战：为我们节省了大量时间和精力。这个服务确实让我们感 觉很周到，现在企业办事确实也更加方便了。 央广网北京6月29日消息 据中央广播电视总台中国之声《新闻和报纸摘要》报道，在深入贯彻中央 八项规定精神学习教育中，重庆渝中、甘肃平凉坚持问题导向，结合工作实际，动真碰硬解决问题。 重庆渝中区楼宇企业密集，针对之前搭建的企业服务平台，对企业一些需求难以第一时间解决企业 办事慢、多头跑等问题，渝中区探索推广一站式对企服务模式，为企业提供帮办、代办服务。 重庆市渝中区委组织部副部长王静：突出深学细悟、笃行实干，建立问题分类管理、分级领办、挂 单销号机制，切实把学习教育成效转化为推动高质量发展、高效能治理、高品质生活的实效。 甘肃平凉聚焦政务办理程序繁杂、材料冗杂等导致群众来回跑路的问题，联动多部门打通堵点，升 级一网通办、一窗受理，为企业和群众办事提效、减负。 甘肃省平凉市委组织部副部长陈涛：坚持立查立行改、集中整治改、为民解难改，整改整治作风顽 瘴痼疾，确保作风建设常态化长效化，以实际行动坚定拥护"两个确立"、坚决做到"两个维护"。 目前，重庆渝中区在学习教育过程中，重点梳理企业实际 ...