CINXE.COM

机器之心 | 企业人工智能服务

<!DOCTYPE html><html lang="en"><head><meta charset="UTF-8" /><meta content="width=device-width, initial-scale=1.0, maximum-scale=1.0, minimum-scale=1.0, user-scalable=no" name="viewport" /><meta content="ie=edge" http-equiv="X-UA-Compatible" /><meta content="telephone=no" name="format-detection" /><title>机器之心 | 企业人工智能服务</title> <meta name="description" content="机器之心"> <meta name="keywords" content="全球人工智能, 人工智能信息服务, 人工智能优质文章, 人工智能媒体, 人工智能产业服务, ai00, ai talk, interface, ai live, 机器智能, 机器学习, 深度学习, 神经网络, 强化学习, 计算机视觉, 自然语言处理, 机器人"> <meta property="og:url" content="https://www.jiqizhixin.com/"> <meta property="og:title" content="企业人工智能服务"> <meta property="og:type" content="website"> <meta property="og:site_name" content="机器之心"> <meta name="twitter:card" content="summary"> <meta name="twitter:site" content="@SyncedTech"> <meta name="twitter:title" content="企业人工智能服务"> <meta name="twitter:description" content="机器之心"> <meta name="twitter:image" content="https://cdn.jiqizhixin.com/assets/global/logo-4819103cf20202b394b95f4d561b26f2959f5be5b58198c02f5a869244beff8c.png"><meta name="csrf-param" content="authenticity_token" /> <meta name="csrf-token" content="8EaAud764hHQQPxAR6Hf67MzcGT+SubaY+4j1hDAptPaXDHSDO1NOns4aHwDm0bXVnEgQZei5vWMJMMpdwuAYg==" /><link rel="stylesheet" media="all" href="https://cdn.jiqizhixin.com/assets/packs/css/application-f55f5e99.css" /></head><body data-controller="home" id="home-index"><header class="header " id="header"><div class="u-container u-flex header__container"><div class="u-flex-center"><a alt="首页" class="header__logo" href="https://www.jiqizhixin.com/"><img alt="机器之心" height="32" src="https://cdn.jiqizhixin.com/assets/logo-black-814ff978059dd2570cc09283d01d1e29ebec941b013fc43d3ae6ce3b3f6c2d69.png" /></a><div class="header-nav__readings"><div class="header-nav__readings--item">报道<div class="header-nav__menu"><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/categories/basic">人工智能</a><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/industry">数字化转型</a><a class="header-nav__menu--item auto-byte">汽车科技<img alt="beta" class="header-nav__menu__logo" src="https://cdn.jiqizhixin.com/assets/auto-byte-icon-fc15735267b7900910ae1da4b36f1c5bce9be4201bf1b5be534dddb2c7ebdaf5.png" /></a><a class="header-nav__menu--item science-ai">交叉前沿<img alt="beta" class="header-nav__menu__logo" src="https://cdn.jiqizhixin.com/assets/science-ai-icon-caa725ee64b502ae7a4e72ed02892e9142f63a50f32772dc960879005e292108.png" /></a></div></div><a class="header-nav__readings--item" href="https://www.jiqizhixin.com/columns">专栏</a><a class="header-nav__readings--item" href="/chat">心思<img alt="new" src="https://cdn.jiqizhixin.com/assets/special_plan/new-5f5f546bf5fa3fcd07522953f6fb6235c7b9eb8ba3f75da72f4a220e6f1fe1b0.svg" /></a><a class="header-nav__readings--item" href="https://app6ca5octe2206.pc.xiaoe-tech.com/index">学堂</a><div class="header-nav__readings--item header-nav__readings--item-sp"><a class="header-nav__readings--item" href="https://www.jiqizhixin.com/short_urls/41640723-0077-46e7-aa2c-c2d67823c0d1" target="_blank">年度榜单2023</a><div class="header-nav__menu header-nav__menu__awards"><p class="header-nav__menu--title">年度榜单</p><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/short_urls/41640723-0077-46e7-aa2c-c2d67823c0d1">年度榜单 2023</a><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/short_urls/e098d7fb-3929-440c-9ecb-0e568106c325">年度榜单 2022</a><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/awards/2021">年度榜单 2021</a><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/awards/2020">年度榜单 2020</a><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/awards/2019">年度榜单 2019</a><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/awards/2018">年度榜单 2018</a><a class="header-nav__menu--item" href="https://mp.weixin.qq.com/s/gw5KgjgK31YdFxlQYGbhsw">年度榜单 2017</a><div class="header-nav__menu--line"></div><p class="header-nav__menu--title">A100 系列</p><a class="header-nav__menu--item" href="https://www.jiqizhixin.com/articles/2020-09-29-9">A100 数智中国</a><p class="header-nav__menu--item header-nav__menu--item--disable">A100 Data Intelligence</p><a class="header-nav__menu--item" href="https://github.com/jiqizhixin/AI00">A100 Original</a></div></div><a class="header-nav__readings--item" href="https://www.jiqizhixin.com/events">活动</a><a class="header-nav__readings--item" href="https://www.jiqizhixin.com/short_urls/1c9524be-37f8-4b87-b6e4-318aae2f7656">SOTA模型</a></div><div class="header-nav__items"><a class="header-nav__item" href="https://pro.jiqizhixin.com/">机器之心 Pro</a></div></div><div class="header-search__ele" id="js-site-search"></div><div class="header__btns"><a class="header-other__link header-user__menu js-switch-sign" href="javascript:;" id="js-header-login">登录</a><a class="header-en-link" href="https://syncedreview.com/">EN</a></div></div></header><div class="code-modal code-modal__auto-byte"><img class="code-modal__close" src="https://cdn.jiqizhixin.com/assets/code-modal__close-c158a550a9982aa92801c046548d72945aa912de6f2a012513132fdc02dfc379.png" /><p class="code-modal__name">Auto Byte</p><p class="code-modal__desc">专注未来出行及智能汽车科技</p><img class="code-modal__img" src="https://cdn.jiqizhixin.com/assets/auto-byte__qr-code-89b58c6d2398c12b7def0f2538cac5ff08727062918f9bad97f25b1d61aac000.png" /><p class="code-modal__title">微信扫一扫获取更多资讯</p></div><div class="code-modal code-modal__science-ai"><img class="code-modal__close" src="https://cdn.jiqizhixin.com/assets/code-modal__close-c158a550a9982aa92801c046548d72945aa912de6f2a012513132fdc02dfc379.png" /><p class="code-modal__name">Science AI</p><p class="code-modal__desc">关注人工智能与其他前沿技术、基础学科的交叉研究与融合发展</p><img class="code-modal__img" src="https://cdn.jiqizhixin.com/assets/science-ai__qr-code-a12e4635c8d3968645bd2b6371981b65b0d21058c5a2c216aef80015b4eb7e43.png" /><p class="code-modal__title">微信扫一扫获取更多资讯</p></div><div class="code-modal__mask"></div><div class="home u-has-header" id="js-has-modal"><div class="u-container"><div class="u-index-section home__top"><div class="home__left"><div class="post swiper-container" id="js-post-slider"><div class="post__content-slider swiper-wrapper"><div class="swiper-slide"><a alt="榜单2024" class="post__cover" href="https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&amp;mid=2650942791&amp;idx=1&amp;sn=955edd97a3ef4efcd7d5ef99d87f397e&amp;chksm=84e7e939b390602f626de0b7ffafccc040e95471f64c295bbc053e51f3c9ed2f28c00a1d4778&amp;token=1900393283&amp;lang=zh_CN&amp;poc_token=HGg3QGej7rbridlCRRep4vKXdThrfl6awN3MBlSL" rel="noopener noreferrer" target="_blank"><img alt="榜单2024" src="https://image.jiqizhixin.com/uploads/external/cover_image/225/%E5%B9%B4%E5%BA%A6%E6%A6%9C%E5%8D%95_-_%E5%AE%98%E7%BD%91%E6%B5%B7%E6%8A%A5_-_Final.png" width="100%" /></a></div><div class="swiper-slide"><a alt="端侧AI" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/course_pc_detail/camp_pro/course_2o97YBORiHu17WPxEHZ7A3YAfVf" rel="noopener noreferrer" target="_blank"><img alt="端侧AI" src="https://image.jiqizhixin.com/uploads/external/cover_image/224/8-5.png" width="100%" /></a></div><div class="swiper-slide"><a alt="Stable Diffusion 实战课" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/goods_pc_detail/goods_detail/course_2go68dUnsjJuFMXH0LHKvJGm8uY" rel="noopener noreferrer" target="_blank"><img alt="Stable Diffusion 实战课" src="https://image.jiqizhixin.com/uploads/external/cover_image/215/%E5%AE%9E%E6%88%98%E8%AF%BE%E7%A8%8B%E5%B0%81%E9%9D%A2%E5%AE%98%E7%BD%91%E4%B8%8A%E7%BA%BF.png" width="100%" /></a></div><div class="swiper-slide"><a alt="视频生成" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/goods_pc_detail/goods_detail/course_2bL2OSsIl4iBgIdbdol9whb2xZC" rel="noopener noreferrer" target="_blank"><img alt="视频生成" src="https://image.jiqizhixin.com/uploads/external/cover_image/211/%E8%A7%86%E9%A2%91%E7%94%9F%E6%88%90.png" width="100%" /></a></div><div class="swiper-slide"><a alt="向量数据库" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/goods_pc_detail/goods_detail/course_2YpgQqltbPYxjXfLGfeFHrR5uy5" rel="noopener noreferrer" target="_blank"><img alt="向量数据库" src="https://image.jiqizhixin.com/uploads/external/cover_image/212/%E5%90%91%E9%87%8F%E6%95%B0%E6%8D%AE%E5%BA%93.png" width="100%" /></a></div><div class="swiper-slide"><a alt="llama2" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/goods_pc_detail/goods_detail/p_64ec66a1e4b04c1014b42ee6?product_id=p_64ec66a1e4b04c1014b42ee6" rel="noopener noreferrer" target="_blank"><img alt="llama2" src="https://image.jiqizhixin.com/uploads/external/cover_image/210/Llama2.png" width="100%" /></a></div><div class="swiper-slide"><a alt="量子经典课" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/goods_pc_detail/goods_detail/p_64ef0260e4b04c1014b4928d?content_app_id=appQTULvsIE4217&amp;product_id=p_64ef0260e4b04c1014b4928d" rel="noopener noreferrer" target="_blank"><img alt="量子经典课" src="https://image.jiqizhixin.com/uploads/external/cover_image/206/WechatIMG3806.jpeg" width="100%" /></a></div><div class="swiper-slide"><a alt="声纹" class="post__cover" href="https://app6ca5octe2206.pc.xiaoe-tech.com/p/t_pc/goods_pc_detail/goods_detail/course_2CmcrDJntdRPxyyKgO2LDn7c2Qe" rel="noopener noreferrer" target="_blank"><img alt="声纹" src="https://image.jiqizhixin.com/uploads/external/cover_image/209/16_9_2x_%E5%89%AF%E6%9C%AC.png" width="100%" /></a></div></div><div class="post__switch"><a class="post__btn post__btn--prev" href="javascript:;"><i class="iconfont icon-prev3"></i></a><div class="post__page"></div><a class="post__btn post__btn--next" href="javascript:;"><i class="iconfont icon-next2"></i></a></div></div></div><div class="home__right home-daily"><div class="home-right__header"><div class="home-right__tab home-right__tab--active home-right__tab-1"><img alt="icon" class="tab-icon" src="https://cdn.jiqizhixin.com/assets/home/tab-1-1413467f550616aaa3cd0eee9b1613e519ede3818041239a1467119e5bd1dde8.png" /><img alt="icon" class="tab-icon--active" src="https://cdn.jiqizhixin.com/assets/home/tab-1--active-46d81d549ae16a35a7b8a42fa823ad30c129e5778c6f5eed55d1292d4cd67e53.png">最新通讯</img></div><div class="home-right__tab home-right__tab-2"><img alt="icon" class="tab-icon" src="https://cdn.jiqizhixin.com/assets/home/tab-2-02744560ebe987fd7b76776abc24ed1680f563545b57aaacd80b7cbd1983c98f.png" /><img alt="icon" class="tab-icon--active" src="https://cdn.jiqizhixin.com/assets/home/tab-2--active-1c2c1721805938ed7591f5d80126ef04642aab1e577a27ff838076887aced2e2.png">最新参考</img></div></div><div class="home-right__body"><div class="home-right__list-1"><div class="home-right__list-item"><div><p>Week 47 · RL「误人」?LeCun 在技术路线上又有何战略摇摆?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.22</span></div><p>本周, NEO 推出可自主学习的 AI 工程师,Kaggle 表现优于 o1;月之暗面推出 k0-math 推理模型以及 Kimi 探索版功能。</p></div><div class="home-right__list-item"><div><p>Week 46 · AI 竞赛进入推理阶段:扩展测试时间计算是万能的吗?Scaling What 成为关键</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.16</span></div><p>本周,外媒曝 OpenAI 调整技术策略引热议;MIT 和哈佛分别发文探讨 Scaling Law 局限与改进机会。</p></div><div class="home-right__list-item"><div><p>Week 45 · 「压缩即智能」,成就 LLM 的 Transformer 未必是终极解?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.08</span></div><p>本周,零一万物宣布战略转向 toB 市场;知乎、昆仑万维、夸克均推出AI搜索功能。</p></div><div class="home-right__list-item"><div><p>Week 44 · 从文本到屏幕:「Project Jarvis」们能实现 AGI 吗?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.01</span></div><p>本周,OpenAI 推出 ChatGPT Search;MIT 新作发现 LLM 可能具有类似人脑的分层结构。</p></div><div class="home-right__list-item"><div><p>Week 43 · 微调失格?持续反向传播算法将解锁新的训练范式吗?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>10.25</span></div><p>本周,Yoshua Bengio 团队提出强化学习新策略;Anthropic 升级版 Claude 3.5 Sonnet 解锁「计算机使用」能力。</p></div><div class="home-right__list-item"><div><p>Week 42 · OpenAI 的 MLE-bench 被当成「奇点触发器」了?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>10.18</span></div><p>本周,OpenAI 开源多智能体框架「Swarm」被指控剽窃;李飞飞提出面向机器人策略训练的「数字表亲」概念。</p></div></div><div class="home-right__list-2"><div class="home-right__list-item"><div><p>RL「误人」?LeCun 在技术路线上又有何战略摇摆?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.22</span></div><div class="desc"><p>Andrey Karpathy 近日有关「悔做 RL」的言论引起许多关注,而他在后悔的同时对 Yann LeCun 从一开始就不看好强化学习的态度始终正确的说法引发了 AI 社区对另一件事的兴趣,即「LeCun 的 AI 战略是否出现过摇摆?」</p></div></div><div class="home-right__list-item"><div><p>Sora 发布大半年后,主流视频生成模型们是否已实现超越?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.22</span></div><div class="desc"><p>在 Sora 发布后的近 9 个月时间里,MiniMax-Video-01、Kling 1.5 等视频生成模型陆续发布,不少表示已经达到了 Sora 级别。</p></div></div><div class="home-right__list-item"><div><p>2025 IT 优先事项报告:在产线中集成 AI 是保持竞争力的关键</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.22</span></div><div class="desc"><p>近日,软件资产管理公司 Flexera 发布《2025 IT 优先事项报告》(以下简称为「报告」)。该报告对来自美国、英国等地区 800 名 IT 领导者进行了调查,重点关注 IT 决策者如何制定 2025 年的优先事项。</p></div></div><div class="home-right__list-item"><div><p>谁能进入下一轮?具身智能「练习生」的技术储备和商业路径有何异同?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.16</span></div><div class="desc"><p>具身智能在 2022 年被李飞飞比喻成视觉领域的一颗「北极星」,而后被认为是通往 AGI 的有效路径。以大语言模型的突破为契机,通过赋予 AI 身体,使其能够和世界交互,再将大脑升级为世界模型,让机器人能够灵活应对未见过的任务,适应工业、居家生活、医疗、养老等不同场景。这样的概念佐以技术进展让具身智能的市场规模不断膨胀,并涌现了大量的创业团队。这些具身智能「练习生」或是在机器人领域长期深耕,具有丰富的技术储备,或者在浪潮中入局,以深厚的学术背景和创新的技术成果吸引大量资本。</p></div></div><div class="home-right__list-item"><div><p>扩展测试时计算是万能的吗?Scaling What成为关键</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.15</span></div><div class="desc"><p>近日,外媒 The Information 的一篇报道引发了关于「模型 Scaling Laws 范式是否已经达到极限」的争议。OpenAI o1 模型的推出将 Scaling Laws 范式从预训练带向了推理层,模型推理层的开发和规模增长成为 AI 大模型玩家们新一轮「竞赛」的重点。涉及复杂计算和决策,缓慢逻辑思考的「System 2」思维成为主导。</p></div></div><div class="home-right__list-item"><div><p>「压缩即智能」,成就LLM的Transformer 未必是终极解?</p><img alt="icon" src="https://cdn.jiqizhixin.com/assets/home/list-icon-89b2748aed60768f98d8e67f934eadfc5449fe6854c06f9e2fe493f87d00f520.png" /><span>11.09</span></div><div class="desc"><p>Ilya Sutskever 2023 年向黄仁勋描述的「AI 预测下一个单词可以实现真正的理解」的理念在一系列 LLM 进展中被验证,其带来的「压缩即智能」的观点进而得到了大量关注和认可。GPT 的训练过程被描述为对知识的无损压缩,但当前最流行的 Transfomer 架构却未必是这一路径下的最优选择。</p></div></div></div></div><div class="home-right__footer">每周获取决策者必须了解的AI&Robotics业内要事</div><div class="home-right__modal-root"><div class="home-right__modal-content"><img alt="icon" class="close" src="https://cdn.jiqizhixin.com/assets/home/modal-close-a09a936d5343e64ada41bd8c1a522f69c9063e2abe0ff564742c427b09d41566.png" /><img alt="icon" class="cover" src="https://www.jiqizhixin.com/short_urls/e03a5197-2cad-4155-baa5-d34c95061a30" /></div><div class="home-right__modal-layer"></div></div></div></div><div class="u-clearfix"><div class="u-col-9 u-block__container js-article-container"><div class="u-block__header"><p class="u-block__title">最新</p><div class="u-block__right"><a alt="全部" class="u-block__tab js-u-tab is-active" data-tab="全部" href="javascript:;">全部</a><a alt="入门" class="u-block__tab js-u-tab" data-tab="basic" href="javascript:;">入门</a><a alt="理论" class="u-block__tab js-u-tab" data-tab="theory" href="javascript:;">理论</a><a alt="工程" class="u-block__tab js-u-tab" data-tab="practice" href="javascript:;">工程</a><a alt="产业" class="u-block__tab js-u-tab" data-tab="industry" href="javascript:;">产业</a></div></div><div class="u-block__body home__newest"><div class="u-block__item js-u-item is-active"><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-8" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="昆仑万维「天工大模型4.0」o1版(Skywork o1)正式启动邀请测试" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/73b7c8f8-2539-479b-a070-d06376c04eb6/%E6%88%AA%E5%B1%8F2024-11-27_17.02.09.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="昆仑万维「天工大模型4.0」o1版(Skywork o1)正式启动邀请测试" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-8" target="_self">昆仑万维「天工大模型4.0」o1版(Skywork o1)正式启动邀请测试</a><p class="u-text-limit--two article-item__summary">Skywork o1是由昆仑万维集团发布的具有慢思考推理能力的系列模型。这是国内第一款中文逻辑推理能力的o1模型。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/24fbf651-67b6-4a30-bb0d-a5486de036de" rel="noopener noreferrer" target="_blank"><img alt="新闻助手" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-1-83f4ab515bbe824688aef66da82d70a493aea5bcfe5c83df85db80a77c1591e9.jpg" /></a><a alt="新闻助手" class="article-item__name" href="/users/24fbf651-67b6-4a30-bb0d-a5486de036de" rel="noopener noreferrer" target="_blank">新闻助手</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 17:02:35 +0800">2024/11/27 17:02</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-8" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-8" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-7" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="刚刚,OpenAI员工的套现机会来了!可通过要约收购向软银出售15亿美元股权" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/4ec16bc5-0d7a-4931-b622-3cf639f7a1e3/iShot_2024-11-27_16.09.46.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="刚刚,OpenAI员工的套现机会来了!可通过要约收购向软银出售15亿美元股权" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-7" target="_self">刚刚,OpenAI员工的套现机会来了!可通过要约收购向软银出售15亿美元股权</a><p class="u-text-limit--two article-item__summary">就在刚刚,CNBC 和路透社等多家媒体报道称,OpenAI 获得了软银 15 亿美元新投资,并且还允许员工通过要约收购(tender offer)出售股权。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/f9b24057-89a6-48bd-874f-fc91d0d8d513" rel="noopener noreferrer" target="_blank"><img alt="AI好好用" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-3-fb31ad35511be5a49c496964c9d2190933420699e44988a3401765d28949a1f8.jpg" /></a><a alt="AI好好用" class="article-item__name" href="/users/f9b24057-89a6-48bd-874f-fc91d0d8d513" rel="noopener noreferrer" target="_blank">AI好好用</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 16:10:00 +0800">2024/11/27 16:10</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-7" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-7" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-6" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="压缩率达10的48次方,实现蛋白序列空间极端压缩,清华EvoAI登Nature子刊" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/1b48810b-bd62-4c15-a38a-55d28f797d42/iShot_2024-11-27_15.55.18.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="压缩率达10的48次方,实现蛋白序列空间极端压缩,清华EvoAI登Nature子刊" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-6" target="_self">压缩率达10的48次方,实现蛋白序列空间极端压缩,清华EvoAI登Nature子刊</a><p class="u-text-limit--two article-item__summary">清华大学的研究团队建立了一种称为 EvoScan 的方法,用于全面分割和扫描高适应度序列空间,以获得能够捕捉其基本特征(尤其是在高维度中)的锚点。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/70d794c4-1662-41a5-8ce6-7ca588c90bdc" rel="noopener noreferrer" target="_blank"><img alt="ScienceAI" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-3-fb31ad35511be5a49c496964c9d2190933420699e44988a3401765d28949a1f8.jpg" /></a><a alt="ScienceAI" class="article-item__name" href="/users/70d794c4-1662-41a5-8ce6-7ca588c90bdc" rel="noopener noreferrer" target="_blank">ScienceAI</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 15:55:00 +0800">2024/11/27 15:55</time></span></div><div class="article-item__other"><a alt="theory" class="u-btn--gray category__link" href="/categories/theory" rel="noopener noreferrer" target="_blank">理论</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-6" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-6" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-5" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="国产大模型首发中文逻辑推理,「天工大模型4.0」o1版来了" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/14b4c700-fed8-4d1a-be5f-d078351b3116/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_202411271304101.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="国产大模型首发中文逻辑推理,「天工大模型4.0」o1版来了" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-5" target="_self">国产大模型首发中文逻辑推理,「天工大模型4.0」o1版来了</a><p class="u-text-limit--two article-item__summary">没想到,技术发展得竟然这么快。最近,人们已经开始畅想 AI 时代后的生活了。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-2-c1597a15e88db92a6875ac5a53546c5a941c23c12158c0c4bd046ee45e9f1ba9.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 13:24:32 +0800">2024/11/27 13:24</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-5" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-5" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-4" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="遗憾不?原来百度2017年就研究过Scaling Law,连Anthropic CEO灵感都来自百度" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/1860bce4-43f2-460a-905b-f6bef9826495/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_20241127130410.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="遗憾不?原来百度2017年就研究过Scaling Law,连Anthropic CEO灵感都来自百度" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-4" target="_self">遗憾不?原来百度2017年就研究过Scaling Law,连Anthropic CEO灵感都来自百度</a><p class="u-text-limit--two article-item__summary">在追求 AGI 的道路上,Scaling Law 是绕不开的一环。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-2-c1597a15e88db92a6875ac5a53546c5a941c23c12158c0c4bd046ee45e9f1ba9.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 13:19:18 +0800">2024/11/27 13:19</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-4" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-4" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-3" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="HuggingFace工程师亲授:如何在Transformer中实现最好的位置编码" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/d86c67ee-2c6a-4ebf-ba5c-7fa2c1a2552f/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_202411271304091.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="HuggingFace工程师亲授:如何在Transformer中实现最好的位置编码" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-3" target="_self">HuggingFace工程师亲授:如何在Transformer中实现最好的位置编码</a><p class="u-text-limit--two article-item__summary">一个有效的复杂系统总是从一个有效的简单系统演化而来的。——John Gall</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-3-fb31ad35511be5a49c496964c9d2190933420699e44988a3401765d28949a1f8.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 13:14:40 +0800">2024/11/27 13:14</time></span></div><div class="article-item__other"><a alt="practice" class="u-btn--gray category__link" href="/categories/practice" rel="noopener noreferrer" target="_blank">工程</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-3" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-3" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27-2" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="跨模态大升级!少量数据高效微调,LLM教会CLIP玩转复杂文本" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/86410c17-2e91-4d3b-86da-5c8c342ec0ef/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_20241127130409.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="跨模态大升级!少量数据高效微调,LLM教会CLIP玩转复杂文本" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27-2" target="_self">跨模态大升级!少量数据高效微调,LLM教会CLIP玩转复杂文本</a><p class="u-text-limit--two article-item__summary">在当今多模态领域,CLIP 模型凭借其卓越的视觉与文本对齐能力,推动了视觉基础模型的发展。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-1-83f4ab515bbe824688aef66da82d70a493aea5bcfe5c83df85db80a77c1591e9.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 13:08:20 +0800">2024/11/27 13:08</time></span></div><div class="article-item__other"><a alt="practice" class="u-btn--gray category__link" href="/categories/practice" rel="noopener noreferrer" target="_blank">工程</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27-2" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27-2" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-27" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="Sora就这么泄露了三小时,网友调侃Altman急拔网线,艺术家们也在抗议被「白嫖」" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/bcb0d85d-21d9-43cd-bd63-dcc14073431c/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_20241127104007.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="Sora就这么泄露了三小时,网友调侃Altman急拔网线,艺术家们也在抗议被「白嫖」" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-27" target="_self">Sora就这么泄露了三小时,网友调侃Altman急拔网线,艺术家们也在抗议被「白嫖」</a><p class="u-text-limit--two article-item__summary">一觉醒来,挺突然的,Sora API 就这么泄露了。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-2-c1597a15e88db92a6875ac5a53546c5a941c23c12158c0c4bd046ee45e9f1ba9.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-27 10:40:41 +0800">2024/11/27 10:40</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-27" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-27" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-26-9" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="Runway和Luma又开始对掐!一手实测全新功能,谁能胜出?" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/b8e2a0c9-533b-4dec-94bf-702ac2dc35e8/0_-_2024-11-26T185940.643.jpeg?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="Runway和Luma又开始对掐!一手实测全新功能,谁能胜出?" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-26-9" target="_self">Runway和Luma又开始对掐!一手实测全新功能,谁能胜出?</a><p class="u-text-limit--two article-item__summary">Runway三天更新两次,Luma全力狙击,AI视频圈狼烟再起。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/f9b24057-89a6-48bd-874f-fc91d0d8d513" rel="noopener noreferrer" target="_blank"><img alt="AI好好用" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-2-c1597a15e88db92a6875ac5a53546c5a941c23c12158c0c4bd046ee45e9f1ba9.jpg" /></a><a alt="AI好好用" class="article-item__name" href="/users/f9b24057-89a6-48bd-874f-fc91d0d8d513" rel="noopener noreferrer" target="_blank">AI好好用</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-26 19:00:14 +0800">2024/11/26 19:00</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-26-9" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-26-9" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-26-8" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="6.85亿次AI加速模拟,分析2万种材料,Meta发布催化剂数据集OCx24" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/f9af6050-bb45-464d-90df-c10c371da4fb/iShot_2024-11-26_16.51.13.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="6.85亿次AI加速模拟,分析2万种材料,Meta发布催化剂数据集OCx24" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-26-8" target="_self">6.85亿次AI加速模拟,分析2万种材料,Meta发布催化剂数据集OCx24</a><p class="u-text-limit--two article-item__summary">近日,Meta FAIR 与多伦多大学和 VSParticle(荷兰纳米技术工程公司 )合作,发布了开放催化剂实验 2024(OCx24)数据集</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/0db0310e-6f96-46bc-9151-a50b36421a76" rel="noopener noreferrer" target="_blank"><img alt="ScienceAI" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-1-83f4ab515bbe824688aef66da82d70a493aea5bcfe5c83df85db80a77c1591e9.jpg" /></a><a alt="ScienceAI" class="article-item__name" href="/users/0db0310e-6f96-46bc-9151-a50b36421a76" rel="noopener noreferrer" target="_blank">ScienceAI</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-26 16:51:00 +0800">2024/11/26 16:51</time></span></div><div class="article-item__other"><a alt="theory" class="u-btn--gray category__link" href="/categories/theory" rel="noopener noreferrer" target="_blank">理论</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-26-8" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-26-8" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-26-6" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="创业一年半,胖了30斤,AI大佬感叹:还是回谷歌好" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/990b5f95-bd3b-423a-bb0e-96b5a415c1a7/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_202411261306473.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="创业一年半,胖了30斤,AI大佬感叹:还是回谷歌好" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-26-6" target="_self">创业一年半,胖了30斤,AI大佬感叹:还是回谷歌好</a><p class="u-text-limit--two article-item__summary">回到大厂,和老领导重聚。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-3-fb31ad35511be5a49c496964c9d2190933420699e44988a3401765d28949a1f8.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-26 13:27:56 +0800">2024/11/26 13:27</time></span></div><div class="article-item__other"><a alt="industry" class="u-btn--gray category__link" href="/industry" rel="noopener noreferrer" target="_blank">产业</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-26-6" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-26-6" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><article class="article-item__container"><a class="js-open-modal" data-remote="" href="/articles/2024-11-26-5" target="_self"><div class="u-opacity-effect u-image-base article-item__cover "><img alt="吴恩达出手,开源最新Python包,一个接口调用OpenAI等模型" class="u-image-center" src="https://image.jiqizhixin.com/uploads/article/cover_image/707798ba-67b8-40b9-a21f-25afbc896803/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_202411261306472.png?imageView2/1/w/390/h/282" /></div></a><main class="article-item__right"><section><a alt="吴恩达出手,开源最新Python包,一个接口调用OpenAI等模型" class="article-item__title t-strong js-open-modal" data-remote="" href="/articles/2024-11-26-5" target="_self">吴恩达出手,开源最新Python包,一个接口调用OpenAI等模型</a><p class="u-text-limit--two article-item__summary">在构建应用程序时,与多个提供商集成很麻烦,现在 aisuite 给解决了。</p></section><footer class="u-flex article-item__footer"><div class="article-item__author"><a class="u-avatar-base article-item__avatar" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank"><img alt="机器之心" class="u-image-center" src="https://cdn.jiqizhixin.com/assets/avatars/anonymous-2-c1597a15e88db92a6875ac5a53546c5a941c23c12158c0c4bd046ee45e9f1ba9.jpg" /></a><a alt="机器之心" class="article-item__name" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61" rel="noopener noreferrer" target="_blank">机器之心</a><span class="article-item__publish"><time class="js-time-ago" datetime="2024-11-26 13:25:17 +0800">2024/11/26 13:25</time></span></div><div class="article-item__other"><a alt="basic" class="u-btn--gray category__link" href="/categories/basic" rel="noopener noreferrer" target="_blank">入门</a><span><a class="u-like-icon article-item__link js-like-action" data-path="/articles/2024-11-26-5" href="javascript:;"><i class="iconfont icon-like-default"></i><span></span></a><a class="u-like-icon article-item__link" href="/articles/2024-11-26-5" rel="noopener noreferrer" target="_blank"><i class="iconfont icon-message-default"></i></a></span></div></footer></main></article><div data-react-class="home/HomeNewest" data-react-props="{&quot;cursor&quot;:&quot;MTI=&quot;,&quot;categoryName&quot;:&quot;all&quot;,&quot;isLoadMore&quot;:true}" data-react-cache-id="home/HomeNewest-0"></div></div><div class="u-block__item js-u-item"><div data-react-class="home/HomeNewest" data-react-props="{&quot;categoryName&quot;:&quot;basic&quot;}" data-react-cache-id="home/HomeNewest-0"></div></div><div class="u-block__item js-u-item"><div data-react-class="home/HomeNewest" data-react-props="{&quot;categoryName&quot;:&quot;theory&quot;}" data-react-cache-id="home/HomeNewest-0"></div></div><div class="u-block__item js-u-item"><div data-react-class="home/HomeNewest" data-react-props="{&quot;categoryName&quot;:&quot;practice&quot;}" data-react-cache-id="home/HomeNewest-0"></div></div><div class="u-block__item js-u-item"><div data-react-class="home/HomeNewest" data-react-props="{&quot;categoryName&quot;:&quot;industry&quot;}" data-react-cache-id="home/HomeNewest-0"></div></div></div></div><div class="u-col-3 has-gutter u-block__container"><div class="u-block__header"><p class="u-block__title">最热文章</p></div><div class="u-block__body hot__container" id="hot-container"><div class="hot-cover"><a class="u-opacity-effect js-open-modal" data-remote="true" href="/articles/2020-09-29-9"></a><img alt="机器之心A100数智中国榜发布:让中小企业看懂、选对、用好数智化服务方案(附研究报告)" class="hot-cover__cover" src="https://image.jiqizhixin.com/uploads/article/cover_image/c8121edf-5d9c-4473-8efa-4d219a1469d0/A100%E6%95%B0%E6%99%BA%E6%A6%9C.png?imageView2/1/w/252/h/160" width="100%" /><div class="hot-cover__content"><a alt="最热文章" class="u-text-limit--two hot-cover__title js-open-modal" data-remote="true" href="/articles/2020-09-29-9">机器之心A100数智中国榜发布:让中小企业看懂、选对、用好数智化服务方案(附研究报告)</a><div class="u-flex hot-item__footer"><a alt="机器之心" class="hot-item__author" href="/users/b6ef5e00-e24f-44ea-9404-78e24df29424"> by 机器之心</a><time>09/29</time></div></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2017-12-27-5">除了深度学习,你需要知道AI技术的23个方向 | 机器之心首份技术报告</a><div class="u-flex hot-item__footer"><a alt="机器之心" class="hot-item__author" href="/users/7f316f0c-8f72-4231-bb30-0eb1dd5a5660">机器之心</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2017-12-27-5#comment"><i class="iconfont icon-message-default"></i>8</a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-22-8">腾讯混元大模型核心论文曝光:Scaling law、MoE、合成数据以及更多 </a><div class="u-flex hot-item__footer"><a alt="新闻助手" class="hot-item__author" href="/users/24fbf651-67b6-4a30-bb0d-a5486de036de">新闻助手</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-22-8#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-21-7">扣子OpenAPI突进智能语音战场!点满低延时、定制化、随时打断和音色克隆技能(内测开启!)</a><div class="u-flex hot-item__footer"><a alt="机器之心" class="hot-item__author" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61">机器之心</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-21-7#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-22-7">无需预训练,亲和力与天然蛋白相当,中国科大的蛋白质从头设计方法登Nature子刊</a><div class="u-flex hot-item__footer"><a alt="ScienceAI" class="hot-item__author" href="/users/a249537c-aa1c-49ea-bd38-e9394ec95be9">ScienceAI</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-22-7#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-21-8">用AI反制AI诈骗,合合信息获全球AI攻防挑战赛金融场景赛道冠军</a><div class="u-flex hot-item__footer"><a alt="机器之心" class="hot-item__author" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61">机器之心</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-21-8#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-22-6">2秒完成检测,将AI与3D打印相结合,科学家实现小型核反应堆高级监测</a><div class="u-flex hot-item__footer"><a alt="ScienceAI" class="hot-item__author" href="/users/a249537c-aa1c-49ea-bd38-e9394ec95be9">ScienceAI</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-22-6#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-22-2">大模型不会推理,为什么也能有思路?有人把原理搞明白了</a><div class="u-flex hot-item__footer"><a alt="机器之心" class="hot-item__author" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61">机器之心</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-22-2#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-21-10">「颠覆性」科学没有衰落,谷歌 AI 推动科学发展的 9 种方式</a><div class="u-flex hot-item__footer"><a alt="ScienceAI" class="hot-item__author" href="/users/a249537c-aa1c-49ea-bd38-e9394ec95be9">ScienceAI</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-21-10#comment"><i class="iconfont icon-message-default"></i></a></div></div><div class="hot-item"><a class="u-text-limit--two hot-item__title js-open-modal" data-remote="true" href="/articles/2024-11-23-5">这才是真・开源模型!公开「后训练」一切,性能超越Llama 3.1 Instruct</a><div class="u-flex hot-item__footer"><a alt="机器之心" class="hot-item__author" href="/users/294c393b-25f7-45b0-bec6-33e3bd344e61">机器之心</a><a alt="评论最热文章" class="u-like-icon hot-item__comment" href="/articles/2024-11-23-5#comment"><i class="iconfont icon-message-default"></i></a></div></div></div></div></div></div></div><div class="modal-layer is-hidden" id="js-modal"><div class="modal-layer__mask"></div><div id="js-modal-container"></div></div><script>document.addEventListener('DOMContentLoaded', (event) => { $('.ncov-ad').length > 0 && $('.ncov-ad')[0].addEventListener('click', function() { ga('create', 'UA-96130205-2', 'auto', 'trackNcovADClickevent'); ga('trackNcovADClickevent.send', 'event', 'ncov_ad_click') }) })</script><div data-user="null" id="js-login-functionality"></div><div class="u-relative"><div class="u-fix-right--list" id="js-fix-right-list"><div class="u-fix-right--btn backtop" id="js-backtop"><a class="backtop__link" href="javascript:;"><i class="iconfont icon-xiangshangjiantou backtop__icon"></i></a><a class="backtop__link is-second">返回顶部</a></div></div></div><footer class="footer" id="footer"><div class="u-flex footer__wrapper"><div class="footer__left"><div class="footer__logo"><img height="43" src="https://cdn.jiqizhixin.com/assets/logo-c617614d41c836153141ce68ff2b8be19e15cd9c16b2ef1936bc4ad734397392.png" /></div><div class="footer-link"><a class="footer-link__item" href="/about" target="_blank">关于我们</a><a class="footer-link__item" href="/terms" target="_blank">服务条款</a></div></div><div class="footer__middle"><div class="footer__title t-second">全球人工智能信息服务</div><div class="footer__middle--items"><span class="footer__middle--friends">友情链接:</span><a alt="Synced Global" class="footer__middle--item" href="https://syncedreview.com/" target="_blank">Synced Global</a><a alt="机器之心 Medium 博客" class="footer__middle--item" href="https://medium.com/@Synced" target="_blank">机器之心 Medium 博客</a><a alt="PaperWeekly" class="footer__middle--item" href="http://www.paperweekly.site/" target="_blank">PaperWeekly</a><a alt="动脉网" class="footer__middle--item" href="https://vcbeat.top/" target="_blank">动脉网</a><a alt="艾耕科技" class="footer__middle--item" href="https://www.aidigger.com/" target="_blank">艾耕科技</a></div></div><div class="footer__right"><div class="footer__social"><div class="footer__tooltip"><a class="footer__social--icon iconfonts icon-wechat" href="javascript:;"></a><div class="footer__tooltip-box"><img alt="机器之心微信公众平台" src="https://cdn.jiqizhixin.com/assets/application/weixinQR-2e2eb9e01d350c1ce6bd7cfe4cf8aa2209a1d72b50f1a6147a371e85548fcdf4.jpg" /></div></div><div class="footer__tooltip"><a class="footer__social--icon iconfonts icon-weibo" href="http://weibo.com/synced" rel="noopener noreferrer" target="_blank"></a></div><div class="footer__tooltip"><a class="footer__social--icon iconfonts icon-rss" href="https://jiqizhixin.com/rss" rel="noopener noreferrer" target="_blank"></a></div></div></div></div><p class="footer__other">©2024 机器之心(北京)科技有限公司 &nbsp;&nbsp;&nbsp;&nbsp;<a class="footer__other" href="https://beian.miit.gov.cn" target="_blank">京ICP备2021005318号-1</a></p></footer><footer class="footer u-flex ncov__footer" id="ncov-footer"><div class="u-flex footer__wrapper"><div class="footer__left"><div class="footer__logo"><img height="50" src="https://cdn.jiqizhixin.com/assets/logo-c617614d41c836153141ce68ff2b8be19e15cd9c16b2ef1936bc4ad734397392.png" /></div><div class="ncov__footer__day"><img src="https://cdn.jiqizhixin.com/assets/weapp/ai_ncov-fd4187ebf31cb21f3d7066c6e8da2202fe1c8c5b4f19ecda437a00d03bae9aa2.svg" /><p>识别二维码,立即订阅智能战疫日报</p></div><div class="ncov__footer__synced">众智成城,共克时艰</div></div><div class="footer__right"><div class="ncov--title">加入「智能战疫联合行动」 </div><div class="ncov__footer__advices"><div class="ncov__footer__advices__item">您/贵司在疫情中面临哪些困难与诉求</div><div class="ncov__footer__advices__item">您/贵司在疫情中可为其他企业提供哪类帮助</div><div class="ncov__footer__advices__item">您/贵司有哪些抗疫动作或故事希望与大家分享</div><div class="ncov__footer__advices__item">您/贵司希望如何参与抗疫行动</div></div><div class="ncov__footer__form"><a href="http://jiqizhixin.mikecrm.com/j2hVkH0">前往填写</a></div></div></div></footer><div data-react-class="utils/Modal" data-react-props="{}" data-react-cache-id="utils/Modal-0"></div><div data-react-class="utils/Alert" data-react-props="{}" data-react-cache-id="utils/Alert-0"></div><div class="notice__container" id="js-notice-container"></div><script src="https://cdnjs.cloudflare.com/ajax/libs/plupload/2.1.6/plupload.full.min.js"></script><script src="https://o.alicdn.com/captcha-frontend/aliyunCaptcha/AliyunCaptcha.js"></script><script>(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){ (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o), m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m) })(window,document,'script','https://www.google-analytics.com/analytics.js','ga'); ga('create', 'UA-96130205-2', 'auto'); ga('send', 'pageview');</script><script src="https://cdn.jiqizhixin.com/assets/packs/js/application-7aa65d627ec44a72b455.js"></script><script>window.currentUser = null; window.sysEnv = "production"</script></body></html>

Pages: 1 2 3 4 5 6 7 8 9 10