鹭羽 发自 凹非寺 量子位 | 公众号 QbitAI u1s1，现在模型能力是Plus了，但Rollout阶段的速度却越来越慢…… 于是月之暗面出手了： 爆改RL训练速度，让LLM"越跑越快"！ 最近月之暗面联合清华大学提出了全新的加速引擎 Seer ，能够在不改变核心训练算法的前提下，大幅度提升LLM的强化学习训练速度。 依托组内上下文设计，可实现同步RL的Rollout效率提升 74%～97% ，长尾延迟减少 75%～93% 。 好好好，几乎是模型换代式的效率提升。 下面来康康详细内容。 跑得更快、更省资源 强化学习目前已成为推动LLM发展的核心技术，但现有系统面临着严重的性能瓶颈。 具体来说，就是在端到端迭代过程中，生成阶段 （rollout phase） 会耗费大量的时间资源，然而该阶段受固有工作负载不均衡的影响，存在 明显的长尾延迟问题，且资源利用率较低。 因此研究团队针对性推出了高效同步RL框架 Seer 。 其核心架构包括三大模块： 1、 推理引擎池 （Inference Engine Pool） 基于DRAM/SSD构建，包括多个推理实例与跨节点的 全局KVCache池 ，不仅可以支持负载均衡 ...

MEET组委会 发自 凹非寺 量子位｜公众号 QbitAI 12月10日 ， 量子位MEET2026智能未来大会 将带你聚焦这一年里最受关注的前沿技术与产业落地进展。 我们将以 「共生无界，智启未来」 为主题，关注以AI为代表的智能科技如何穿透产业、学科与场景的边界，成为驱动社会演进的核心动能。 强化学习、多模态、芯片算力、AI+行业、AI出海 等等今 年科技圈最热议的话题，你都能够在这场大会上看到。 这里既有 学术前沿 与 商业落地 的最新碰撞，也有来自 Infra 、 模型 、 产品产业 的领先技术成果。 大会上还将权威发布 人工智能年度榜单 与 年度AI趋势报告 ，敬请期待。 话不多说，现在大会已经开启了 观众报名通道 ，点击链接线下参会 今年MEET智能未来大会依然盛况不减，第一二波嘉宾阵容在此， 一起来看有哪些大咖嘉宾出席—— 最新嘉宾阵容 张亚勤 清华大学智能产业研究院院长 中国工程院院士 张亚勤院士于2014年9月至2019年10月担任百度公司总裁。出任百度总裁前，张亚勤院士曾在微软公司工作16年，历任全球资深副总裁兼微软亚太研发集团主 席、微软亚洲研究院院长兼首席科学家、微软全球副总裁和微软中 ...

除了教职和创业者身份外，苏昊还是最早提出"Embodied AI"概念的学者之一；高阳曾在UC伯克利深耕机器视觉研究，拥有具身大模型训练 与机器人真实场景实验上的丰富一线经验。 面对具身智能具身智能研究及产业化落地，就技术演化、模型路线等议题，两人有共识，也有不同的观点。 对谈重点围绕几个核心问题展开： 中国具身智能两位先锋的转折与抉择 具身是否是通往AGI的必经之路？ 技术突破的真正瓶颈在哪里？ 中国在这条路径上的结构性优势是什么？ 下一次可感知的跃迁会在何时到来？ 允中 发自 凹非寺 量子位 | 公众号 QbitAI 两位长期站在具身智能第一线的亲历者，给出了罕见清晰的判断。 UCSD终身教授、Hillbot联合创始人苏昊 直截了当，表示"没有具身智能就没有通用物理智能、通用智能"。 清华大学助理教授、千寻智能的联合创始人高阳 则补上一句现实的验证路径："只要去scale数据就能解决，本质没有任何区别。" 绿洲资本举办的AGM （Annual General Meeting） 现场，就具身智能这个话题——它正从一个技术分支，转变为理解通用智能的关键入 口，苏昊和高阳分享了自己的洞察。 绿洲 ：想先请两位老 ...

机器之心报道 机器之心编辑部 刚刚，人工智能顶会 NeurIPS 2025 公布了最佳论文奖、时间检验奖等奖项！ 今年共有 4 篇论文获得最佳论文奖，另有 3 篇论文获得最佳论文亚军（Best Paper Runner-up）。 这七篇论文的研究涵盖了多个前沿方向，包括：扩散模型理论、自监督强化学习、大语言模型中的注意力机制、LLM 的推理能力、在线学习理论等。 另外，任少卿、何恺明、Ross Girshick、孙剑 2015 年合著论文《Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks》获得了时间检验 奖。 NeurIPS 还宣布了Sejnowski-Hinton 奖的获奖者：获奖论文是《Random synaptic feedback weights support error backpropagation for deep learning》。 本届 NeurIPS 会议共收到 21575 份有效投稿并进入评审流程，最终接收 5290 篇，整体录用率为 24.52%。 以下是获奖论文的详细信息： ...

点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近30个 方向 学习 路线 端到端作为这两年的量产关键词，是各家车企核心的招聘岗位。但市面上真正的量产人才少之又少，从模型优化、场景优化、数据优化，再到下游的规划兜底，端 到端其实是一个全栈的岗位。从技术的成熟度和工业界的需求来看，端到端需要攻克的难题还有很多。导航信息的引入、强化学习调优、轨迹的建模及优化都有很 多门道，目前也是量产第一线。 为此我们花了三个月的时间设计了端到端量产进阶课程，从实战到落地层层展开。 该课程涉及的核心算法包括：一段式端到端、两段式端到端、导航信息的量产应用、开闭环强化学习、扩散模型+强化学习、自回归+强化学习、时空联合规划等 等，最后分享一些实际的量产经验。很多想进阶或者跳槽的同学苦于没有专家辅导，想转行但实际工作中无法接触到实际的量产优化，简历上往往不够亮眼，遇到 问题连个请教的人都没有。 这门课程是自动驾驶之心联合工业界算法专家开设的《面向量产的端到端实战小班课》！课程只有一个重点：聚焦量产。从一段式、两段式、强化学习、导航应 用、轨迹优化、兜底方案再到具体量产经验分享。面向就业直击落地，所以这门课程 ...

具身智能之心技术交流群成立了！主要关注VLA、VLN、遥操作、Diffusion Policy、强化学习、VLA+RL、 sim2real、多模态大模型、仿真、运动控制、目标导航、建图定位、导航等方向。 感兴趣的同学可以添加小助理微信AIDriver005，邀请加入我们的社群。 注意哦， 备注：机构/学校+姓名+研究方向 ，能够快速入群！ ...

Core Insights - The MEET2026 Smart Future Conference will focus on cutting-edge technologies and industry developments that have garnered significant attention throughout the year [1] - The theme "Symbiosis Without Boundaries, Intelligence to Ignite the Future" emphasizes how AI and smart technologies are penetrating various industries, disciplines, and scenarios, becoming a core driving force for societal evolution [2] Group 1: Conference Highlights - The conference will cover hot topics in the tech circle this year, including reinforcement learning, multimodal AI, chip computing power, AI in various industries, and AI going global [3] - The event will showcase the latest collisions between academic frontiers and commercial applications, featuring leading technological achievements from infrastructure, models, and product industries [4] - The conference will also feature the authoritative release of the annual AI rankings and the annual AI trend report [5][93] Group 2: Notable Speakers - Zhang Yaqin, President of Tsinghua University's Intelligent Industry Research Institute and an academician of the Chinese Academy of Engineering, has a notable background in AI and digital video technologies [11][12] - Sun Maosong, Executive Vice President of Tsinghua University's AI Research Institute, has led multiple national projects and has extensive experience in AI research [15] - Wang Zhongyuan, Director of the Beijing Academy of Artificial Intelligence, has a strong background in AI core technology development and has published over 100 papers [19] Group 3: AI Trends and Rankings - The "Artificial Intelligence Annual Rankings" initiated by Quantum Bit has become one of the most influential rankings in the AI industry, evaluating companies, products, and individuals across three dimensions [94] - The "2025 Annual AI Trend Report" will analyze ten major AI trends based on technological maturity, implementation status, and potential value, highlighting representative institutions and best cases [95] Group 4: Event Details - The MEET2026 Smart Future Conference is scheduled for December 10, 2025, at the Beijing Jinmao Renaissance Hotel, with registration now open [96] - The conference aims to attract thousands of tech professionals and millions of online viewers, establishing itself as an annual barometer for the smart technology industry [98]

Core Insights - Ilya Sutskever, co-founder of OpenAI and a key figure in deep learning, has shifted focus from scaling models to research-driven approaches in AI development [1][2][3] - The industry is moving away from "scale-driven" methods back to "research-driven" strategies, emphasizing the importance of asking the right questions and developing new methodologies [2][3] - Sutskever argues that while AI companies may experience stagnation, they can still generate significant revenue despite reduced innovation [2][3] - The potential for narrow AI models to excel in specific domains suggests that breakthroughs may come from improved learning methods rather than merely increasing model size [3][4] - The emergence of powerful AI could lead to transformative societal changes, including increased productivity and shifts in political and governance structures [3][4] - Sutskever emphasizes the importance of aesthetic principles in research, advocating for simplicity and elegance in AI design [4] Industry Trends - The scaling laws that dominated AI development are nearing their limits, prompting a return to foundational research and exploration [2][28] - The current phase of AI development is characterized by a shift from pre-training to reinforcement learning, which is more resource-intensive [29][30] - The distinction between effective resource utilization and mere computational waste is becoming increasingly blurred in AI research [30][31] - The scale of computational resources available today is substantial, but the focus should be on how effectively these resources are utilized for meaningful research [42][44] Company Insights - Safe Superintelligence (SSI) has raised $3 billion, positioning itself to focus on foundational research without the pressures of market competition [45][46] - SSI's approach to AI development may differ from other companies that prioritize immediate market applications, suggesting a long-term vision for advanced AI [45][46] - The company believes that the true value lies not in the sheer amount of computational power but in the strategic application of that power to drive research [43][44]

MEET组委会 发自 凹非寺 量子位｜公众号 QbitAI 2025年，我们正迈入一个由人工智能重塑一切的新时代。 12月10日，量子位MEET2026智能未来大会 将带你聚焦这一年里最受关注的前沿技术与产业落地进展。 我们将以 「共生无界，智启未来」 为主题，关注以AI为代表的智能科技如何穿透产业、学科与场景的边界，成为驱动社会演进的核心动能。 强化学习、多模态、芯片算力、AI+行业、AI出海 等等今 年科技圈最热议的话题，你都能够在这场大会上看到。 这里既有 学术前沿 与 商业落地 的最新碰撞，也有来自 Infra 、 模型 、 产品产业 的领先技术成果。 大会上还将权威发布 人工智能年度榜单 与 年度AI趋势报告 ，敬请期待。 话不多说，现在大会已经开启了 观众报名通道 ，点击链接线下参会 今年MEET智能未来大会依然盛况不减，最新嘉宾阵容在此， 一起来看还有哪些大咖嘉宾出席—— 张亚勤 清华大学智能产业研究院院长 中国工程院院士 张亚勤院士于2014年9月至2019年10月担任百度公司总裁。出任百度总裁前，张亚勤院士曾在微软公司工作16年，历任全球资深副总裁兼微软亚太研发集团主 席、微软亚洲研究院院长 ...

Core Viewpoint - The article emphasizes the need for adaptability and continuous learning in the investment landscape, particularly in the context of emerging technologies like AI and biotechnology, highlighting the importance of maintaining a growth mindset amidst uncertainty [6][7][11]. Group 1: Investment Landscape - The current investment environment is characterized by collective anxiety within the Chinese venture capital community, questioning how to navigate a landscape devoid of simple innovation models [7]. - The transition from traditional investment strategies to hard technology sectors, such as biotechnology, poses significant challenges for seasoned investors who must adapt to new paradigms [9][10]. - The concept of "infinite games" is introduced, suggesting that successful companies focus on continuous evolution rather than short-term victories, which is crucial for long-term sustainability [24][25]. Group 2: Cultural Confidence - There is a deep-rooted cultural confidence in Chinese entrepreneurship, reflecting a historical resilience and a spirit of innovation that persists despite challenges [12][13]. - The belief in a new cycle of innovation, termed "Innovation 2.0," is gaining traction among investors and entrepreneurs, indicating a shift towards optimism in the market [12][16]. Group 3: AI and Future Trends - The emergence of AI is seen as a transformative force that will redefine productivity, enabling individuals and small teams to achieve significant market valuations [17]. - The article discusses the potential for AI to integrate into various industries, suggesting that its true impact will be realized when it becomes ubiquitous in decision-making processes [17][19]. Group 4: Narrative and Collaboration - The ability to create compelling narratives is highlighted as a unique human trait that drives collaboration and innovation, essential for achieving extraordinary outcomes [19][20]. - Successful companies are described as those that not only provide solutions but also construct an attractive vision for the future, fostering a shared sense of purpose among stakeholders [21][22]. Group 5: Learning and Growth - Continuous learning and iteration are emphasized as critical components of success in an ever-evolving business landscape, with failures viewed as valuable learning experiences [28][30]. - The article concludes with a call for entrepreneurs to embrace challenges and maintain a commitment to growth, underscoring that great companies thrive by remaining engaged in the market and evolving over time [30].