| 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 ...

Core Viewpoint - The article discusses the post-training strategies of Vision-Language-Action (VLA) models from the perspective of human motor skill learning, emphasizing the need for robots to undergo a post-training phase to adapt to specific tasks and environments, similar to how humans learn skills through practice and experience [4][5][9]. Summary by Sections 1. Introduction to VLA Models - VLA models integrate visual perception, language understanding, and action generation, enabling robots to interact with their environment effectively. However, their out-of-the-box performance is often insufficient for complex real-world applications, necessitating a post-training phase to refine their capabilities [8][9]. 2. Post-Training Strategies - The article categorizes VLA model post-training strategies into three dimensions: environment perception, embodiment (body awareness), and task understanding. This classification mirrors the key components of human motor learning, facilitating targeted improvements in specific model capabilities [10][12]. 3. Environmental Perception Enhancement - Strategies include enhancing the model's ability to perceive and adapt to various operational environments, utilizing cues from the surroundings to inform actions, and optimizing visual encoding for task-specific scenarios [12][13]. 4. Body Awareness and Control - The post-training strategies focus on developing internal models that predict body state changes, improving the model's ability to control robotic movements through feedback mechanisms inspired by human motor control [14]. 5. Task Understanding and Planning - The article highlights the importance of breaking down complex tasks into manageable steps, akin to human learning processes, to enhance the model's understanding of task objectives and improve operational planning [14]. 6. Multi-Component Integration - Effective skill acquisition in humans involves synchronizing multiple learning components. Similarly, VLA models benefit from integrating various strategies to optimize performance across different dimensions [14]. 7. Challenges and Future Trends - Despite advancements, challenges remain in enabling robots to learn and adapt like humans. Key areas for future research include improving kinematic models, optimizing action output structures, and enhancing human-robot interaction through expert knowledge integration [16][17][18]. 8. Continuous Learning and Generalization - The need for continuous learning capabilities is emphasized, as current VLA models often struggle with retaining previously learned skills. Future research should focus on developing algorithms that allow for lifelong learning and better generalization in open environments [22]. 9. Safety and Explainability - The article underscores the importance of safety and explainability in robotic decision-making, advocating for research into interpretable AI and safety mechanisms to ensure reliable operation in diverse scenarios [22].

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 ...

Group 1 - The meeting emphasized the importance of the United Front Work as a core responsibility of the Party Committee, highlighting the need to deeply understand Xi Jinping's thoughts on the United Front Work in the new era [1] - The meeting discussed the necessity to implement the responsibility system for United Front Work, focusing on key tasks such as industrial and technological innovation, reform breakthroughs, private economy development, and optimizing the business environment [1] - The meeting underscored the unique advantages of the United Front in terms of intellectual resources, talent, and broad connections, aiming to promote high-quality development of the United Front Work in the new era [1] Group 2 - The meeting reviewed the work report of the Municipal Commission for Discipline Inspection and Supervision for the first half of the year, stressing the ongoing severe and complex situation of the anti-corruption struggle [2] - The meeting highlighted the need for a combination of strict management and care, addressing issues of inaction and misconduct while also supporting responsible officials through mechanisms for error correction and clarification [2] - The meeting discussed the implementation of the Central Eight Regulations and emphasized the importance of integrating strict requirements into educational efforts, focusing on addressing prominent issues in work style [2] Group 3 - The meeting also reviewed the work related to the compilation of the "14th Five-Year Plan" and discussed other matters [3]

Core Points - The meeting emphasized the importance of young cadres in learning and education, highlighting the need to understand the significance of the Central Committee's deployment and the long-term effectiveness of the Eight Regulations [1][2] - Zhao Yide stressed the necessity for young cadres to enhance their political qualities and to draw nourishment from Xi Jinping's thoughts, improving their political judgment, comprehension, and execution [2] Group 1 - Young cadres are the focus of the learning education, and they must recognize the serious nature of the "Four Winds" issues and implement educational tasks thoroughly [1] - The process of refining work style should be seen as a way to strengthen party character and enhance political awareness [2] - Young cadres should actively engage in key tasks and reforms, demonstrating a spirit of responsibility and problem-solving [2] Group 2 - The meeting called for a strategic focus on the cultivation of young cadres, emphasizing the need for strict education, management, and support [2] - It was highlighted that young cadres should maintain high moral standards and be vigilant against corruption, emphasizing the importance of family education and integrity [2]

Core Insights - The article outlines an eight-step roadmap for efficiently advancing in AI/ML by focusing on essential skills and avoiding common pitfalls [1]. Group 1: Step-by-Step Learning Path - **Step 1: Master Python Core Libraries** Proficiency in Python is essential for AI/ML, including data cleaning, model building, and result visualization [2]. Key content includes Python basics, advanced AI programming techniques, and libraries like scikit-learn, NumPy, Matplotlib, Seaborn, and Pandas [4]. Recommended resources include CS50 Python course and "Python Data Science Handbook" [4]. Suggested learning period is 3-4 weeks [4]. - **Step 2: Solidify Mathematical Foundations** A strong grasp of linear algebra, probability, and calculus is crucial for understanding models [5]. Key content includes matrix operations, Bayesian thinking, and optimization techniques [5]. Recommended resources include "Linear Algebra" by 3Blue1Brown and MIT's Probability Introduction [5]. Suggested learning period is 4-6 weeks [5]. - **Step 3: Understand Machine Learning Basics** This step is pivotal for transitioning from beginner to competent AI/ML engineer [6]. Key content includes supervised vs. unsupervised learning, reinforcement learning, and deep learning [6]. Recommended resources include Google's Machine Learning Crash Course and "Machine Learning" by Andrew Ng [8]. Suggested learning period is 6-8 weeks [8]. - **Step 4: Hands-On Project Experience** Practical experience through real AI/ML applications is essential for job readiness [9]. Key content includes practical guides and project development [9]. Suggested learning period is ongoing [9]. - **Step 5: Learn MLOps** Understanding MLOps is vital for deploying and maintaining models in real-world scenarios [10]. Key content includes foundational concepts and best practices for model deployment [10]. Suggested learning period is 3-4 weeks [10]. - **Step 6: Specialize in a Domain** After building a foundation, focusing on a specific area like NLP or computer vision enhances employability [11]. Suggested learning period is ongoing [11]. - **Step 7: Stay Updated** Continuous learning is necessary to keep skills relevant in the fast-evolving AI field [12]. Key resources include ArXiv for research papers and notable figures in the field [12]. Suggested learning period is ongoing [12]. - **Step 8: Prepare for Interviews** Comprehensive preparation for interviews is crucial, including explaining model principles and system design [13]. Recommended resources include machine learning interview guides [13]. Suggested learning period is 4-6 weeks [13]. Conclusion - The article emphasizes a structured approach to mastering AI/ML, enabling individuals to transition from novices to job-ready professionals efficiently [1].

"这哪行？暴雨来了，肯定得涝！"姜旺对张青云说，"必须抓紧按照排水渠的标准，挖深、拓宽、延 长。" "村里一时拿不出这么多钱……"张青云说。 "得要这么深，才不怕暴雨。"跳进排水渠，姜旺用身体作尺，量了量渠深。 "放心！这渠深1米、宽80厘米。"一旁的黄狮村党支部书记张青云说。 抢夏收、忙夏播、抓防汛，这段时间，河南省南阳市西峡县五里桥镇党委书记姜旺奔忙在田间地 头，"现在会议少了，有更多时间抓具体工作了。" 深入贯彻中央八项规定精神学习教育开展以来，西峡县积极推进减会议、减报表、优调研，让干部腾出 更多时间和精力，到一线办实事。 以精简会议为例，西峡县推行"多会合一"，将同类型的会议套开。以往，县农业农村局、水利局、交通 运输局、住房和城乡建设局都要召开防汛会议，今年全县安全生产、抗旱、防汛3个会议套开。今年以 来，西峡县召开各类会议数量同比下降35%，会议时长缩减50%以上。 减了会议，多出来的时间用在哪里？西峡县要求干部——"不能只在空调房里听汇报，必须直插基层一 线，去解决实际问题。" 来到黄狮村的猕猴桃园，姜旺发现，路旁只有一条灌溉用的水沟。 刚在这处检查完，又有村民找过来："修渠把路给弄断了，我们 ...

会议要求，市政府党组要切实扛牢主体责任，各级领导干部要坚持以身作则、以上率 下，带头入脑入心学、全面彻底查、注重实效改，全力营造风清气正、干事创业的良好政治 生态。 编辑：胡之澜 6月27日，市政府召开党组（扩大）会议，传达学习中央第三指导组指导督导湖北见面 会精神、省委党的建设工作领导小组会议暨深入贯彻中央八项规定精神学习教育推进会精 神，部署政府系统贯彻落实措施。市委副书记、市长、市政府党组书记盛阅春主持会议并讲 话。 会议指出，在全党开展深入贯彻中央八项规定精神学习教育，是今年党建工作的重点任 务。全市政府系统要切实把思想和行动统一到党中央决策部署和省委、市委工作要求上来， 进一步增强抓好学习教育的政治自觉、思想自觉、行动自觉，把接受指导督导作为践行"两 个维护"的政治检验、纵深推进全面从严治党的有力抓手，全力支持配合中央指导组开展工 作，确保学习教育取得实效。 会议强调，要在一体推进学查改上再加力再深化，对中央指导组指出的问题照单全收、 立行立改，确保学有质量、查有力度、改有成效。要以推动解决突出问题为重点，结合巡视 巡察、整治形式主义为基层减负、整治群众身边不正之风和腐败问题等重点工作，明确查摆 问 ...