<p>新的一年到来了，过去的 2016 年可以说是有史以来机器学习领域进展最显著的一年。在大数据和高性能计算设备的助力下，具备学习能力的机器在围棋、语音识别、翻译、图像渲染和识别等许多领域都实现了惊人的成就。但那远远还不是这一领域的终点，大部分媒体和投资者仍然还看好以机器学习为主的人工智能技术的未来发展，市场也无法掩饰地表现出了对这方面的人才的需求。</p>    <p>这篇文章尽可能地全面地梳理了机器之心在 2016 年发过的基础知识和技术指导方面的文章，希望能为读者通往人工智能领域的专业人才乃至学界大牛之路提供一点助力。本文按照从基础到前沿划分对文章进行了分类（学习资源、基础介绍文章、技术起点、继续进阶、前沿研究），读者可方便地根据自己的学习进度选择合适的文章阅读。</p>    <p>一、学习资源</p>    <p><img src="https://simg.open-open.com/show/99c15a8578ecd60193d53148a2e6affc.jpg"></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721853&idx=3&sn=f9f0048cccefbf9c00dc94f2a71c7d00&chksm=871b0a43b06c8355376d1bc897b4f3585b4a129d77a9e5b025f4c344f2ef264c05133cff7682&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习资料大全：从基础到各种网络 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719733&idx=1&sn=2fce6d18e8fcae9b805d4652a7c702e9&chksm=871b018bb06c889d339f34b192579f2e7f97e2f7b64cbe3ccd85ec7b3ed9022178ed5a150359&scene=21#wechat_redirect" rel="nofollow,noindex">30个深度学习库：按Python和C++等10种语言分类 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721418&idx=1&sn=71b28bce48b70f9fa30929db29e685e1&chksm=871b08f4b06c81e24d16ef66265e142de1fd8a18fd872da99a2e70c4e10062ac41575ec0e382&scene=21#wechat_redirect" rel="nofollow,noindex">2016年不可错过的21个深度学习视频、教程和课程 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721817&idx=4&sn=a2e9f4c19f92fd95040f4b13935c6681&chksm=871b0a67b06c837108989bc0cd0b2dc06ce2c0009190f1ab6cfd38bca5f6d08d6b56e776d811&scene=21#wechat_redirect" rel="nofollow,noindex">程序员实用深度学习免费课程:从入门到实践  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721817&idx=3&sn=f03d9c31c8db61c7d980b0137c9e2da3&chksm=871b0a67b06c8371ddaa7a49f56ee7b5b9725f52e0a3a930b9d424e7d6fae0ca24bca7e319af&scene=21#wechat_redirect" rel="nofollow,noindex">9本不容错过的深度学习和神经网络书籍  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719393&idx=1&sn=41ed306d26dd209acfd61ee70efc8cf6&chksm=871b00dfb06c89c9dc8d5da87a0be1b3666c439909b41559097e952b7ccec29bae09823521bf&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习专业名词表：从激活函数到word2vec   </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721717&idx=2&sn=6a58289f9f65448339a845f88f168088&chksm=871b09cbb06c80dd7b790d34e3627eaf3f330b8670350ee5ad3718bc36e3d5f3f30581fe8453&scene=21#wechat_redirect" rel="nofollow,noindex">2016年年度十大Python库盘点 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650716071&idx=1&sn=7aa209732425c6a52536fbb9012a09fd&scene=21#wechat_redirect" rel="nofollow,noindex">2010-2016年被引用次数最多的深度学习论文  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722147&idx=1&sn=0584e654c66f694502ff2918dcbc3ca1&chksm=871b0b1db06c820bc8396e6b4ab1ced5f0bada892385eabed7fb581c5b911162bfe326bc36a0&scene=21#wechat_redirect" rel="nofollow,noindex">自学数据科学&机器学习，19个数学和统计学公开课推荐  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721121&idx=5&sn=5af0fb9465ce9345c017adfb1d0c9796&chksm=871b0f1fb06c8609b8b9d792b222b16f7b6fa4479a6d426663ab03cf1ebebd5209080bebe48b&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio新书《Deep Learning》中文版开放预览（附PDF下载链接）  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715120&idx=2&sn=108bc0f1bac2deb0b423a2587ebe306a&scene=21#wechat_redirect" rel="nofollow,noindex">数据科学家应该掌握的12种机器学习算法（附信息图） </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720472&idx=1&sn=a0d8f835a300fb5d0c4a3a224cc17924&chksm=871b0ca6b06c85b0461163c3043bf45ab6d58dd57263ce381646ba71842608785d97e00014f6&scene=21#wechat_redirect" rel="nofollow,noindex">从入门到研究，人工智能领域最值得一读的20份资料 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715104&idx=3&sn=2f3c9a625519f6265a3ba755697cad56&scene=21#wechat_redirect" rel="nofollow,noindex">关于数据科学的十本好书  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719875&idx=1&sn=f1030f9dbd1d8080585b4d3bba3a9b73&chksm=871b02fdb06c8beb2dc94e3f69a76fa56aafb3c310d03e53f3aca2167a27a9e115fbb6b0bfe2&scene=21#wechat_redirect" rel="nofollow,noindex">哈佛大学九大自然语言处理开源项目  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718688&idx=1&sn=c473fd7fe002fe6b8b5b917c432daf27&scene=21#wechat_redirect" rel="nofollow,noindex">ICML 2016演讲视频：数百个演讲带你读懂机器学习 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721717&idx=1&sn=470fef6589c81afa913b2e0572996f92&chksm=871b09cbb06c80ddc4c21fc55511b30c6e1a0461c9f1c0588ce9dd484d6216d819e949a1a259&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio研究生科研指导演讲：解读人工智能全貌和下一个前沿  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720663&idx=2&sn=d06c7a2e8cdde84901ffd8335a775737&chksm=871b0de9b06c84ff467a63042b17ad1622224fc1ab84597ab80b64ab0049c4f4b09cabc3d7dc&scene=21#wechat_redirect" rel="nofollow,noindex">Yann LeCun演讲 ：人工 智能的下一个前沿——无监督学习 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720780&idx=1&sn=39ac663849c90ab31d3dd04013f7e646&chksm=871b0e72b06c8764becc21ecb10ece2b7d285f72a85cec79920967abe53bdbf7a16522623caf&scene=21#wechat_redirect" rel="nofollow,noindex">今年GitHub排名前20的Python机器学习开源项目 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715438&idx=3&sn=2c2a6e4dfda2307d818dcab385ef6727&scene=21#wechat_redirect" rel="nofollow,noindex">机器学习工程师和数据科学家最应该读的16本书  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722365&idx=4&sn=af6b4a85b4e447d0d54e71399f6d8e93&chksm=871b1443b06c9d55f1678dbe89573c1fcb31758b12b6b992cb63be11b7b1dc6b9f1874cee0bc&scene=21#wechat_redirect" rel="nofollow,noindex">任何阶段的学习者都适用的参考：机器学习领域书目全集  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718578&idx=1&sn=ff7d748b149e7952c9fa3b53cefd5afc&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio深度学习暑期班学习总结，35个授课视频全部开放 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722318&idx=2&sn=f03f56fb91bfd5fe393268f74435349f&chksm=871b1470b06c9d661a72d727c76e7f75a8cf57c17c97542cc1c8b330b8969b1fd2748722fb95&scene=21#wechat_redirect" rel="nofollow,noindex">Andrej Karpathy CS294课程总结：可视化和理解深度神经网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719828&idx=2&sn=c6a73c885c3a8db5f2b0484f7c20498a&chksm=871b022ab06c8b3c8e93ff797be2d68d96945f01b617584e79593e16dcd4a76b19a4642b441d&scene=21#wechat_redirect" rel="nofollow,noindex">Geoffrey Hinton最新演讲梳理：从人工神经网络到RNN应用 </a></p>    <p>吴恩达NIPS 2016演讲现场直击：如何使用深度学习开发人工智能应用</p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721406&idx=1&sn=8251dfebd3c360d180339c34c4710fa7&chksm=871b0800b06c81160244fcda78d7816da8ec31d5fbec546ba6e36241e6d32c0ca943ce9ce73d&scene=21#wechat_redirect" rel="nofollow,noindex">NIPS 2016最全盘点：主题详解、前沿论文及下载资源 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719377&idx=1&sn=114e536fc2ca22edcb6ad6ceb332228e&chksm=871b00efb06c89f953e6d7320a7f5fed54e062687425c1dee8166b16704fef23f4eb53b1a473&scene=21#wechat_redirect" rel="nofollow,noindex">斯坦福大学周末学习盛宴：12位大牛解读深度学习 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719367&idx=1&sn=1cd2e133a56eb3e841e98ab837086df8&chksm=871b00f9b06c89eff783040ad6b30152440e628908fc478b9b8cbce991f914f94237e1333fb1&scene=21#wechat_redirect" rel="nofollow,noindex">提升深度学习模型的表现，你需要这20个技巧 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720407&idx=1&sn=768d7248e0ab5fa469dbae86d11152e1&chksm=871b0ce9b06c85ffefa2e0c8f6fb7ae4cc1c0500cda7bad008fe68ed6b87d7c8765d138e1fd1&scene=21#wechat_redirect" rel="nofollow,noindex">TensorFlow开源一周年：这可能是一份最完整的盘点  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718160&idx=3&sn=a2114e3324f28740d2603da26fbbdfb4&scene=21#wechat_redirect" rel="nofollow,noindex">五大主流深度学习框架比较分析：MXNET是最好选择  </a></p>    <p>二、基础介绍文章</p>    <p><img src="https://simg.open-open.com/show/2cc2daf5c4e4dc15c06e858549d565ee.jpg"></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719723&idx=3&sn=0338d470c8668aa912aeebc18c1ced0e&chksm=871b0195b06c8883bab39807043824dcec576180f006226672da649459e7c2de0ecd3fd98c6e&scene=21#wechat_redirect" rel="nofollow,noindex">2016伦敦深度学习峰会观感：人工智能面临的三大难题 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722085&idx=1&sn=14a9cc3610e0de25587a707f554939f0&chksm=871b0b5bb06c824da313d70525cca67c9c7417951174455f1e27d9431af132e914c5ab4f6be5&scene=21#wechat_redirect" rel="nofollow,noindex">2016机器学习与自然语言处理学术全景图：卡耐基梅隆大学排名第一  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720407&idx=3&sn=f8484e3aa14bfe9b925d75bb9acc52e2&chksm=871b0ce9b06c85ff78b2ec6a708e0a9128027a8e862a08a2fca38a17d1aa117f48b27402f133&scene=21#wechat_redirect" rel="nofollow,noindex">2016年美国机器人路线图出炉，最新机器人产业盘点  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720758&idx=3&sn=0d7e70cfd3624296fc8aac4944a359dc&chksm=871b0d88b06c849e71808681107f2b695233bdd37d7a79249cad5b2721f4285ed50577bee89d&scene=21#wechat_redirect" rel="nofollow,noindex">一张图看懂全球Bot布局  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715580&idx=1&sn=cc67a90fe35702732f1a67e3e59d80f6&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio：深度学习崛起带来人工智能的春天  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718943&idx=1&sn=258117d392ca1bfc37d6496992da5eae&scene=21#wechat_redirect" rel="nofollow,noindex">神经网络架构演进史：全面回顾从LeNet5到ENet十余种架构 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719010&idx=1&sn=aaa7cc47f27129bbced25e6d090e2c1d&scene=21#wechat_redirect" rel="nofollow,noindex">Andrej Karpathy：计算机科学博士的生存指南  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721559&idx=3&sn=98a74c48397f8a8d9166a3bf0b74fb4f&chksm=871b0969b06c807f90e72c890c4b23308501f46c7315823c3fee1c423817ed81626d3da9d711&scene=21#wechat_redirect" rel="nofollow,noindex">贝叶斯神经网络简史 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720407&idx=2&sn=86603b0276580e9e969e78cac606fe3c&chksm=871b0ce9b06c85ff79c95b4b42ad8bfa8fa14438266e689e7ffe3a882ee024bfd3e0ac90655e&scene=21#wechat_redirect" rel="nofollow,noindex">百度首席科学家吴恩达刊文：人工智能的能力和不足  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721378&idx=1&sn=16f8a955e1459926dd1e66f82e26028c&chksm=871b081cb06c810aedab8a7c1902daf18f3d00f20c1b04ec0645303ffe2023b5e4dd87c41ee8&scene=21#wechat_redirect" rel="nofollow,noindex">高盛百页人工智能生态报告：美国仍是主导力量，中国正高速成长 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720663&idx=1&sn=ffadef6ce7d0e2d60e09c0c8bdc366c5&chksm=871b0de9b06c84ffad7621ec0b3e7578d085d9557f029067d978a286e22ff5e67e0b98ed209b&scene=21#wechat_redirect" rel="nofollow,noindex">伯克利教授Stuart Russell：人工智能基础概念与34个误区 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720174&idx=2&sn=cb195f4dfda478510479eba5dc53fc99&chksm=871b03d0b06c8ac674234d2bcea8515c2720bb502bcb281024b71ebaf4864e1e6fad863d3ebd&scene=21#wechat_redirect" rel="nofollow,noindex">初学者必读：解读14个深度学习关键词 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720335&idx=3&sn=7552135b320b24ce9b273423187b5a78&chksm=871b0c31b06c852753124d9962aff43c049f2283dcfa78cc8ca51289aa0dd2db53d88b2c2e96&scene=21#wechat_redirect" rel="nofollow,noindex">智能时代每个人都应该了解：什么是深度学习？ </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720553&idx=2&sn=8f77792079e14219da4006247d738652&chksm=871b0d57b06c84414a164a210ac350e6fc22af6ba06c2604c749776431fef53731c7f40439c7&scene=21#wechat_redirect" rel="nofollow,noindex">CMU机器学习系负责人：人工智能与人类的未来是共生自主  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715759&idx=2&sn=d005ca89ba3b0bd2eae4b7677cbb8be2&scene=21#wechat_redirect" rel="nofollow,noindex">CMU教授邢波：人工智能的路径、方向与未来  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650716455&idx=2&sn=48da4b101ef293de0f627aeb3ba3231b&scene=21#wechat_redirect" rel="nofollow,noindex">从供应链优化到差异化定价：机器学习十种方式变革制造业  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721938&idx=1&sn=a7568453296f8b34f1bbb6287cb994d0&chksm=871b0aecb06c83fad3b111c2f48206b26c85762e2b9834a07dddd7a282b60a5df2f78c104de5&scene=21#wechat_redirect" rel="nofollow,noindex">对比深度学习十大框架：TensorFlow最流行但并不是最好  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721480&idx=3&sn=0ee6a881f01d8d1bee16d2042ac93e5f&chksm=871b08b6b06c81a027a5b5490f1c9decf3d70002e1f1e1869c25c7a12657f9cfa518b855cf9b&scene=21#wechat_redirect" rel="nofollow,noindex">当AI遇上AR ——从微软HPU说起  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720922&idx=5&sn=110bf38376c42b55d95521dd04944828&chksm=871b0ee4b06c87f2315db196b47620853d3587c69f9486cf5ae9a19c5aa2d4fa101f4541888b&scene=21#wechat_redirect" rel="nofollow,noindex">洞悉AlphaGo超越围棋大师的力量：机器之心邀你一起强化学习 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719022&idx=2&sn=6efb98d3650bd963122744b14b09e1d9&scene=21#wechat_redirect" rel="nofollow,noindex">东南大学漆桂林教授：知识图谱不仅是一项技术，更是一项工程  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717374&idx=1&sn=14c7af07c4d6859a7cb8d7a8650d5825&scene=21#wechat_redirect" rel="nofollow,noindex">人工智能全局概览：通用智能的当前困境和未来可能  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722118&idx=1&sn=1934dfd96eedb777b506a7f9fca1c44a&chksm=871b0b38b06c822e4419d7263ece28aa7ad683c4478eb908397c2fae2aa7505529f9708c72d4&scene=21#wechat_redirect" rel="nofollow,noindex">访谈百度IDL林元庆：百度大脑如何在人脸识别上战胜人类「最强大脑」  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720302&idx=4&sn=5a54ac9d2097f0db11d1da8a65cf2e4c&chksm=871b0c50b06c85466eb0bb5e455e306deca82662d83895f9d6ad2311539a9adf7056dac306ac&scene=21#wechat_redirect" rel="nofollow,noindex">TensorFlow 生态系统：与多种开源框架的融合 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719511&idx=2&sn=ef5f5c1d8d70cc81cdcd2c6fde36f695&chksm=871b0169b06c887f46d4b87987bd58881875f352c3bdb5d8c2e39d9898b6bcac8b7c0c31783d&scene=21#wechat_redirect" rel="nofollow,noindex">Kaggle创始人问答：深度学习会淘汰其他的机器学习方法吗？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721755&idx=1&sn=61837dc0127c1c829522e1772959a0dd&chksm=871b09a5b06c80b3d89848a140f13519bf2c3e264724a8604603751079f43252365e5f40dadd&scene=21#wechat_redirect" rel="nofollow,noindex">机器之心年度盘点 | 从技术角度，回顾2016年语音识别的发展</a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719333&idx=2&sn=ba2491a4c22da7512d9add55c1ac50ee&chksm=871b001bb06c890d71a688cd127557913694240b6a81ea1681552677d63cec27fcd45570dc4d&scene=21#wechat_redirect" rel="nofollow,noindex">机器学习的基本局限性：从一个数学脑筋急转弯说起  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721777&idx=3&sn=20cf64536919147639b2f5cbad38df0f&chksm=871b098fb06c80995100a9afee640203b14f5429f72ee9259b01b6c87d72396beeecdea99144&scene=21#wechat_redirect" rel="nofollow,noindex">人们都在说人工智能，其实现在我们真正做的是智能增强 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717606&idx=4&sn=b94b58d4fe75c1a1e42274720a269a99&scene=21#wechat_redirect" rel="nofollow,noindex">人工智能、机器学习、深度学习，三者之间的同心圆关系  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722126&idx=1&sn=00b5c6ed7e4c576c8be3eae1dc348cfe&chksm=871b0b30b06c8226b163198b78ad40b52715509d3b36f0774670d6e0e284378cc3bdeca42d4a&scene=21#wechat_redirect" rel="nofollow,noindex">R vs Python：R是现在最好的数据科学语言吗？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722239&idx=3&sn=aa556d22c8dc71195175930fda1655d0&chksm=871b0bc1b06c82d70e62081d87c1e75291ea15f631aff5c0a5d03de41ffbbcc4be650e63b014&scene=21#wechat_redirect" rel="nofollow,noindex">斯坦福NLP团队介绍交互式语言学习：从语言游戏到日程规划  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720597&idx=2&sn=815b3426c4855e446e4aa118abdee0d6&chksm=871b0d2bb06c843da11f16f85510da8cb398542c82a80a8822ac9f78beb24aaf58f066eaaed2&scene=21#wechat_redirect" rel="nofollow,noindex">斯坦福大学副教授Reza Zadeh：神经网络越深就越难优化 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=402228099&idx=1&sn=a8e664d332f7d28250fbbf357c773f62&scene=21#wechat_redirect" rel="nofollow,noindex">神经网络和深度学习简史（三）：强化学习与递归神经网络 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=402552632&idx=1&sn=694a4a327a79c4efeeb4db15b3ff4a28&scene=21#wechat_redirect" rel="nofollow,noindex">神经网络和深度学习简史（四）：深度学习终迎伟大复兴  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717977&idx=2&sn=705d47688adadcdce6e09a81e3381e2d&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习技术在股票交易上的应用研究调查 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721378&idx=4&sn=cb49818959281909d8878bea5b34e837&chksm=871b081cb06c810a0075a475beeecbbe41a915679f05d629aa2e4d3a38da45e0899a764a98be&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习十大飙升趋势  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=401846516&idx=1&sn=ed76da1e8f99c604957c8889692de884&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习入门，以及它在物联网和智慧城市中的角色</a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720765&idx=3&sn=a9ec5e451cc7e46ab254b663ec78c3b6&chksm=871b0d83b06c8495b977ba6f82c3aa5974a6d2066e3fec63f9fc33e7e0766a2008804795895e&scene=21#wechat_redirect" rel="nofollow,noindex">服务器端人工智能，FPGA和GPU到底谁更强？ </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721102&idx=1&sn=347d3a3b5fd136df7bc51fb67948fd30&chksm=871b0f30b06c86263fd2d51665aaa5e81d1e3cbc0044ed71ab0f74238a461ba0e75490a04af7&scene=21#wechat_redirect" rel="nofollow,noindex">Science「机器人子刊」创刊号，五大研究解读机器人领域最新进展  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718975&idx=1&sn=2b0ccf0c746e6f10707e5357168d51d6&scene=21#wechat_redirect" rel="nofollow,noindex">西红柿还是猕猴桃？一个案例帮你入门机器学习 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715201&idx=1&sn=d7ad8c79bf060875b6247d634c52449b&scene=21#wechat_redirect" rel="nofollow,noindex">详细解读神经网络十大误解，再也不会弄错它的工作原理  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719646&idx=2&sn=3f84e8af0ca03476eb843f57b497465e&chksm=871b01e0b06c88f6c77601db4a1d224a5007ba306eac7d0522c0b973ca9c3d57c4fabf57a5c4&scene=21#wechat_redirect" rel="nofollow,noindex">基于图的机器学习技术：谷歌众多产品和服务背后的智能 </a>  </p>    <p>三、技术起点</p>    <p><img src="https://simg.open-open.com/show/73aa624ba67a3359f6ed6c3761e2eb39.jpg"></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719723&idx=4&sn=f86e2f30bcee67a424d4617b72560b8d&chksm=871b0195b06c88832108944311c2494a30483590c8c301b0076e2190734a3271e2929db312ed&scene=21#wechat_redirect" rel="nofollow,noindex">10种深度学习算法的TensorFlow实现 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719794&idx=2&sn=8e320caa1d32f7a444a17bfa93b318ad&chksm=871b024cb06c8b5a421efa2eedff5b7149d17028d99e95a38857785db06c1eb2b7796d6e7e4f&scene=21#wechat_redirect" rel="nofollow,noindex">数十种TensorFlow实现案例汇集：代码+笔记  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720367&idx=2&sn=d71e1e6550faece1cb5fb58bd7c919d9&chksm=871b0c11b06c8507d765a781b3ff49c1755deba7f12f36763b468aa23dfafccd768f59538004&scene=21#wechat_redirect" rel="nofollow,noindex">机器学习入门必备：如何用Python从头实现感知器算法 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718527&idx=1&sn=04db4fc59cc23c079a17573657d2b1c7&scene=21#wechat_redirect" rel="nofollow,noindex">ACM 最新月刊文章：强化学习的复兴 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720174&idx=4&sn=da197d35bf9b5ac59b9ffb8e442264f0&chksm=871b03d0b06c8ac622e5ff2b83c786b7fb34d49b6186f1d9f1964b28e31c1e0d7b2ac3863afc&scene=21#wechat_redirect" rel="nofollow,noindex">人工智能开发者的入门指南 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717691&idx=2&sn=3f0b66aa9706aae1a30b01309aa0214c&scene=21#wechat_redirect" rel="nofollow,noindex">从入门到精通：卷积神经网络初学者指南 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718466&idx=1&sn=016f111001e8354d49dd4ce279d283cd&scene=21#wechat_redirect" rel="nofollow,noindex">机器学习敲门砖：任何人都能看懂的TensorFlow介绍  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720463&idx=2&sn=50d88169778ec31a1e1e2d801325005d&chksm=871b0cb1b06c85a7e0299fc733b67d3107970f1176186a2ac8ca2dd5a23d896b5041e592ab4d&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind提出的可微神经计算机架构的TensorFlow实现  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718229&idx=4&sn=f47990a661e1522a5794d6334a334830&scene=21#wechat_redirect" rel="nofollow,noindex">RNN 怎么用？给初学者的小教程 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=401970614&idx=2&sn=479370f70613e431d35c752139c0b602&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习教程：从感知器到深层网络 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721559&idx=4&sn=950d87934d39712fab6164d3d2a37be5&chksm=871b0969b06c807f47b6fac20daae25d6d2f57495f6624554550fad7177f56657d208f40690d&scene=21#wechat_redirect" rel="nofollow,noindex">OpenAI 的 PixelCNN++实现：基于 Python的实现  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719828&idx=2&sn=c6a73c885c3a8db5f2b0484f7c20498a&chksm=871b022ab06c8b3c8e93ff797be2d68d96945f01b617584e79593e16dcd4a76b19a4642b441d&scene=21#wechat_redirect" rel="nofollow,noindex">Geoffrey Hinton最新演讲梳理：从人工神经网络到RNN应用 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717557&idx=1&sn=6c8239ec01c2a3f0def744063d070eec&scene=21#wechat_redirect" rel="nofollow,noindex">官方指南：如何通过玩TensorFlow来理解神经网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720174&idx=1&sn=d770ab6ccbf25c6569d002472daf1b3b&chksm=871b03d0b06c8ac6c16a6e37e1dcc2a5f490e4ab5a21b82b76a2a41d3a1d35d76a6a3d21fdc3&scene=21#wechat_redirect" rel="nofollow,noindex">一篇文章带你进入无监督学习:从基本概念到四种实现模型 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721979&idx=2&sn=146a34534414fb4842398cac62cd201a&chksm=871b0ac5b06c83d3eeb71d5b39d74f9d5648d7e334039cf6d6a3d42b18856abc47ba229b8890&scene=21#wechat_redirect" rel="nofollow,noindex">NIPS 2016上22篇论文的实现汇集  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721979&idx=1&sn=82ccf18462e569d43d72736aef57177a&chksm=871b0ac5b06c83d30e65fcd1344bf4b9773281af6cd58f4d3e6709f24d66f6c2bbbfaddb8ccf&scene=21#wechat_redirect" rel="nofollow,noindex">机器学习初学者入门实践：怎样轻松创造高精度分类网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718717&idx=1&sn=85038d7c906c135120a8e1a2f7e565ad&scene=21#wechat_redirect" rel="nofollow,noindex">解决真实世界问题：如何在不平衡类上使用机器学习？ </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717628&idx=1&sn=a6f1b35f8168f1bc842ca36bb3a69368&scene=21#wechat_redirect" rel="nofollow,noindex">卷积神经网络架构详解：它与神经网络有何不同？ </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717709&idx=2&sn=2bff1e56bc75d65e178476ea9a93b2c5&scene=21#wechat_redirect" rel="nofollow,noindex">LSTM和递归网络基础教程 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718717&idx=2&sn=9cf62b1b684f5dea3c735d1413a50689&scene=21#wechat_redirect" rel="nofollow,noindex">门外汉如何使用谷歌的Prediction API做机器学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718597&idx=2&sn=98c141c6d73eb62a0f3f8aa2a8231b66&scene=21#wechat_redirect" rel="nofollow,noindex">NVIDIA趣味解读：深度学习训练和推理有何不同？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722126&idx=4&sn=96b2fa1fe3fc858d415ddddf2ac3f8e5&chksm=871b0b30b06c82262fec886bc76b0427542dcf8b1fb4bae62d851d032e4fc22d247e981f263f&scene=21#wechat_redirect" rel="nofollow,noindex">如何利用 Python 打造一款简易版 AlphaGo  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715104&idx=2&sn=f6ba338c02f8e08c7821b4ecd5a527e9&scene=21#wechat_redirect" rel="nofollow,noindex">如何用图像识别技术来变革商业？这里有份操作指南  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719093&idx=4&sn=d852318f6b3adb5ef7730def66fb0a93&scene=21#wechat_redirect" rel="nofollow,noindex">MIT生成视频模型，预测静态图片的未来场景  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720257&idx=1&sn=92ac1424c5880a406c1a00558359792b&chksm=871b0c7fb06c8569e23927623aa3c98af8a3377069ba6cd11a0acecb54cd8ce7896e29238b97&scene=21#wechat_redirect" rel="nofollow,noindex">EMNLP 2016干货：从原理到代码全面剖析可用于NLP的神经网络 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718504&idx=1&sn=ba884ec5bd9dd93b77c85f91d8371056&scene=21#wechat_redirect" rel="nofollow,noindex">使用机器学习翻译语言：神经网络和seq2seq为何效果非凡？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720758&idx=1&sn=3004c425e0d427f4900a182d74bed31d&chksm=871b0d88b06c849e951469ae1ed54e5f66074d6322eb6681c85727bb8199154709c04c48c034&scene=21#wechat_redirect" rel="nofollow,noindex">神经网络快速入门：什么是多层感知器和反向传播？ </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718853&idx=4&sn=4f33d503f1602e608b34c6cda4e55dad&scene=21#wechat_redirect" rel="nofollow,noindex">神经网络中激活函数的作用  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720582&idx=2&sn=06076c800c912622d8fa42665d7de1fc&chksm=871b0d38b06c842e3f6edaa5d36046b6702d26a0bcfb710c2010e5f0815389bf4f60077441ef&scene=21#wechat_redirect" rel="nofollow,noindex">逐层剖析，谷歌机器翻译突破背后的神经网络架构是怎样的？ </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719294&idx=1&sn=f1a01cd6710e6ea9629619cd3324d102&chksm=871b0040b06c895642ff961a6fe81f05c5e9776aff5da4845f2d3d874f88213863afd2059833&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习漫游指南：强化学习概览 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717969&idx=1&sn=712e4880e63db42bcb4db5ba06c9856d&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习与神经网络全局概览：核心技术的发展历程  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=402261356&idx=1&sn=f66ee62b002b8a9879d3c428f846e440&scene=21#wechat_redirect" rel="nofollow,noindex">最全的深度学习硬件指南  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719118&idx=2&sn=fad8b7cad70cc6a227f88ae07a89db66&scene=21#wechat_redirect" rel="nofollow,noindex">主流深度学习框架对比：看你最适合哪一款？  </a></p>    <p>四、继续进阶</p>    <p><img src="https://simg.open-open.com/show/c83640110d4706f6d0632eb71123e5d3.jpg"></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720132&idx=1&sn=d630d47c4ab60d35752aba74a9d53361&chksm=871b03fab06c8aec767776a6a4a407c3897dcad26392b24a22536261565e9dc6b5ce52df0816&scene=21#wechat_redirect" rel="nofollow,noindex">40年认知架构研究概览：实现通用人工智能的道路上我们已走了多远？ </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721682&idx=1&sn=6bdbf5739bb312449cb60cb6679f98d2&chksm=871b09ecb06c80fa59dba741fb79e44021d5ae16f67488c38b3fb477235a203931da86c829bc&scene=21#wechat_redirect" rel="nofollow,noindex">第四范式联合创始人陈雨强：机器学习在工业应用中的新思考  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719971&idx=2&sn=c7e0d1f6dd4e9ddce291e9bc2c85c65f&chksm=871b029db06c8b8b7557095989dd3fdb57b86a1d7923c388ca1e74255d07f08992bb0461d958&scene=21#wechat_redirect" rel="nofollow,noindex">词嵌入系列博客Part1：基于语言建模的词嵌入模型  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720050&idx=2&sn=9fedc937d3128462c478ef7911e77687&chksm=871b034cb06c8a5a8db8a10f708c81025fc62084d871ac5d184bab5098cb64e939c1c23a7369&scene=21#wechat_redirect" rel="nofollow,noindex">词嵌入系列博客Part2：比较语言建模中近似 softmax 的几种方法 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720074&idx=2&sn=183fc6285835a48ae7c6bbcce228b063&chksm=871b0334b06c8a22b072f61d4f914210468db7df36a1c6586bd9b6bf3fc6d9f821101d5254c0&scene=21#wechat_redirect" rel="nofollow,noindex">词嵌入系列博客Part3：word2vec 的秘密配方  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718853&idx=1&sn=17462cfef179876db1f9d29dbd95ba2c&scene=21#wechat_redirect" rel="nofollow,noindex">从分割到识别，全面解析非死book开源的3款机器视觉工具 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718786&idx=1&sn=58e6cdf2ffb0f47eb831e0cd623ce0e1&scene=21#wechat_redirect" rel="nofollow,noindex">从硬件到软件：OpenAI 解读自家的深度学习基础架构 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719694&idx=1&sn=93c5aa0b6dd9cdb35c8b8186ce70afff&chksm=871b01b0b06c88a6ee9c3264e67b3e3256f007c4046fefb1d003f91ada71d00547f91b48ca21&scene=21#wechat_redirect" rel="nofollow,noindex">分布式深度学习：神经网络的分布式训练 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720463&idx=4&sn=ddffc71e49002f731ada533cf05cc84d&chksm=871b0cb1b06c85a73aeb3562836c2ff15357417ce7106ceed06f8987b1cdb70a990895592e9d&scene=21#wechat_redirect" rel="nofollow,noindex">Embedding 新框架模型：Exponential Family Embeddings  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719294&idx=3&sn=d23a6d7c03732f7218f70447336801bd&chksm=871b0040b06c89564722f7292215997f474374d43d26761700f00b240e83a9c5dc2dd7595a26&scene=21#wechat_redirect" rel="nofollow,noindex">FPGA vs. ASIC，谁将引领移动端人工智能潮流？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717977&idx=3&sn=e07366137aab6694c3d3d4a311ed6c54&scene=21#wechat_redirect" rel="nofollow,noindex">概述性论文：卷积神经网络的近期研究进展  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715279&idx=2&sn=fd25ed0539b7bf8e79f5e99eea47b889&scene=21#wechat_redirect" rel="nofollow,noindex">《Nature》 封面文章：人工智能引发材料科学变革  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721922&idx=3&sn=b7af3daa477955e0466c521fd45a23e6&chksm=871b0afcb06c83ead80f3cc7e2e650ba441d4068ecd10532272fc4c0b88aadf8b9b24fa8cc19&scene=21#wechat_redirect" rel="nofollow,noindex">概述论文：迁移学习研究全貌  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721602&idx=2&sn=f18e2d3a23dec485350611651e571031&chksm=871b093cb06c802aecc953e10c6bf5a14784ce3bad2c82170262ed6d6d7daedfcea9cceb804d&scene=21#wechat_redirect" rel="nofollow,noindex">Andrej Karpathy：你为什么应该理解反向传播  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721284&idx=1&sn=427e7f45c8253ab22a3960978409f5d1&chksm=871b087ab06c816c424ad03810be3e1b3aa9d6e99a5f325047796f110d178a07736f667d1a10&scene=21#wechat_redirect" rel="nofollow,noindex">GAN之父NIPS 2016演讲现场直击：全方位解读生成对抗网络的原理及未来 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719228&idx=2&sn=b3ccd8c77c2ef81369c02b85de013038&chksm=871b0782b06c8e94e88ce927f8357a6637b051c9f2ace9dbadea84c77b1b8ca0fec7efdb7448&scene=21#wechat_redirect" rel="nofollow,noindex">Google Brain 讲解注意力模型和增强RNN  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719723&idx=2&sn=03925a0eb5a8b78cc2642781b924032c&chksm=871b0195b06c888336f97ac330524580589bf29b073aa76585319a6ad43744a5697d6a424b0f&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌Magenta项目是如何教神经网络编写音乐的？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721644&idx=4&sn=51ff444c29797ea83f55f47c694b2e84&chksm=871b0912b06c8004a9d5cf461a496b693f5c5010896a0b32f1ae5ede9ab32ce2de4b2a8f37b0&scene=21#wechat_redirect" rel="nofollow,noindex">基于MXNet 的神经机器翻译实现  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721121&idx=3&sn=e21d84cb34b75b744a31fb27c3af8528&chksm=871b0f1fb06c8609de11c76854c1ad9582aece4b79eed6d94b588498109d78a2c62ffe69e02c&scene=21#wechat_redirect" rel="nofollow,noindex">2016深度学习重大进展：从无监督学习到生成对抗网络 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715145&idx=1&sn=84ddd1cbb981e260e49100ec39d01663&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习遇上基因组，诊断疾病和揭示深层生物原理或迎来突破  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719794&idx=3&sn=a64c739eab79f95c85685db7b06d3649&chksm=871b024cb06c8b5aa3e8b3ddd59de64dcf32f05b068bda1fa6c9400a491de6f63b1ca4a81d14&scene=21#wechat_redirect" rel="nofollow,noindex">King+Woman-Man=Queen:用基于Spark的机器学习来捕捉词意 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721817&idx=2&sn=110962b592985cb01216feafe0d8510e&chksm=871b0a67b06c8371b96cff1675cc762f4677dc02a94d985c21c28ba8af1c1afccca64080e41f&scene=21#wechat_redirect" rel="nofollow,noindex">初学者必读:从迭代的五个层面理解机器学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722553&idx=5&sn=81228e1becc1895699fd5a87e120be05&chksm=871b1487b06c9d915136b64c14800f60af13ac956060716b654b1a8944beffe0947b036cce47&scene=21#wechat_redirect" rel="nofollow,noindex">轻量级Matlab深度学习框架LightNet的实现  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650714956&idx=3&sn=de8326724feb96cd5891e6198226a365&scene=21#wechat_redirect" rel="nofollow,noindex">如何基于机器学习设计一套智能交易系统？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718229&idx=1&sn=adaf68f2c193028f1c4de5b8d3217cca&scene=21#wechat_redirect" rel="nofollow,noindex">如何在TensorFlow中用深度学习修复图像？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720420&idx=2&sn=179fa42fafe685265fef3b88f186fd62&chksm=871b0cdab06c85cc1500ab14605fe2848188bb55cebd0cfdb1af6c6b1e0c1d2f2b3cf886394a&scene=21#wechat_redirect" rel="nofollow,noindex">机器学习中的并行计算：GPU、CUDA和实际应用  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720922&idx=5&sn=110bf38376c42b55d95521dd04944828&chksm=871b0ee4b06c87f2315db196b47620853d3587c69f9486cf5ae9a19c5aa2d4fa101f4541888b&scene=21#wechat_redirect" rel="nofollow,noindex">深度解读AlphaGo胜利背后的力量：强化学习 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718429&idx=1&sn=46214968459af95e85efe12b8a26b11b&scene=21#wechat_redirect" rel="nofollow,noindex">英伟达自动驾驶技术解读：用于自动驾驶汽车的端到端深度学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720663&idx=3&sn=d9f671f77be23a148d1830448154a545&chksm=871b0de9b06c84ffaf260b9ba2a010108cca62d5ce3dcbd8c98c72c9f786f9cd460b27b496ca&scene=21#wechat_redirect" rel="nofollow,noindex">深度解读最流行的优化算法：梯度下降  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718391&idx=1&sn=99fd9d942768706e93f7f1aa744ea4b7&scene=21#wechat_redirect" rel="nofollow,noindex">Science：斯坦福大学用迁移学习预测非洲贫困状况  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720823&idx=4&sn=2ed3964e94e3076e060e48a4708faa2a&chksm=871b0e49b06c875fd2ead27692e8c2b35ac38724adc0215183a434283f4f869fde11e7036c5e&scene=21#wechat_redirect" rel="nofollow,noindex">用 Word2vec 轻松处理新金融风控场景中的文本类数据  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720969&idx=1&sn=d5723c205f594616932624684d3a1327&chksm=871b0eb7b06c87a1d4d9809b35c361a4d8a4bf75315eaa9502a051058b184fb233e5b2d2695c&scene=21#wechat_redirect" rel="nofollow,noindex">Science：实用量子计算机已近在咫尺  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720561&idx=3&sn=76741dd8ae09d6493b3abebcbd387520&chksm=871b0d4fb06c8459752dff98c4ffd59a8fd6114f1c43b2e1e9c85294402f3adf7947f74e1bab&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习硬件架构简述 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718634&idx=1&sn=1220e691541c34281c64655a01793cb0&scene=21#wechat_redirect" rel="nofollow,noindex">深度学习系列Part2：迁移学习和微调深度卷积神经网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719170&idx=1&sn=68b6b7f87677f5287b6e5a306409653b&chksm=871b07bcb06c8eaa0a649d7d3fd7963423dd4ea51b6e7711bc63653a528fbf196566345ae064&scene=21#wechat_redirect" rel="nofollow,noindex">图文并茂的神经网络架构大盘点：从基本原理到衍生关系 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717143&idx=1&sn=4eb48040935380a7c87d18efea403d58&scene=21#wechat_redirect" rel="nofollow,noindex">为你的深度学习任务挑选性价比最高GPU  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719511&idx=3&sn=0674ef2a2e995d180ca081ed3a6ae1b9&chksm=871b0169b06c887feefc5604f1f53081e919eadec5c3d02c25d8da786de0bfa23141bd3f184e&scene=21#wechat_redirect" rel="nofollow,noindex">详解谷歌神经网络图像压缩技术：高质量地将图像压缩得更小  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718229&idx=3&sn=bb2fd16046e6e08bea612a5f7fd0f2dd&scene=21#wechat_redirect" rel="nofollow,noindex">用于视觉任务的CNN为何能在听觉任务上取得成功？  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722208&idx=1&sn=52397806416c7d7f570d5c8fc9ecb96e&chksm=871b0bdeb06c82c85c03e7a07a3c71d9258969470ed8b70eeff850db98a0b7b98cda6fe787ee&scene=21#wechat_redirect" rel="nofollow,noindex">自然语言处理领域深度学习研究总结：从基本概念到前沿成果  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720780&idx=3&sn=bac6cbe4972c236ee0bcdbf76139fa98&chksm=871b0e72b06c8764837fdf6f6cc8b01361b883c0a86bb253163cde5f380beb2e413d855d84b9&scene=21#wechat_redirect" rel="nofollow,noindex">专访谷歌Jeff Dean：强化学习适合的任务与产品化应用 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720302&idx=1&sn=c88634da158f36db23b9dc7d0dc550ad&chksm=871b0c50b06c854694984e193f289deb51a5efe71f53223dc37feb70509fd957c8af5bb61ab3&scene=21#wechat_redirect" rel="nofollow,noindex">重磅论文：解析深度卷积神经网络的14种设计模式 </a></p>    <p>五、前沿研究</p>    <p><img src="https://simg.open-open.com/show/48a2b686082042d6ab5328563c0f1ea0.jpg"></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721777&idx=4&sn=a55d6807a8e65059c426876c623f655b&chksm=871b098fb06c80990a07affb5ecae55b496ca5e3a8b42e686d287b3bd308235c04a2d29c9a28&scene=21#wechat_redirect" rel="nofollow,noindex">并行运算，非死book提出门控卷积神经网络的语言建模 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718798&idx=3&sn=6ec32a0f0f09b8f193c578ee0af9d7ae&scene=21#wechat_redirect" rel="nofollow,noindex">FAIR与微软研究院合著论文：通过虚拟问答衡量机器智能  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717728&idx=2&sn=228e5905d7cd6fd175fc596e2d511eed&scene=21#wechat_redirect" rel="nofollow,noindex">FusionNet融合三个卷积网络：识别对象从二维升级到三维  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720234&idx=3&sn=1fe8cce38750ad900e31109e4358ee0a&chksm=871b0394b06c8a826c9af5302dc38ccc8279db3288b573227d0547d27a720aec5dba2592e0e6&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌新论文提出神经符号机：使用弱监督在Freebase上学习语义解析器  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720399&idx=4&sn=3dcdb45229af883cd646985fe0964537&chksm=871b0cf1b06c85e7eeebdb857bcc95b456e5f5ccdb597c9739d6b2fda96069bc1fc53cb32b6b&scene=21#wechat_redirect" rel="nofollow,noindex">Google Brain与OpenAI合作论文：规模化的对抗机器学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721644&idx=5&sn=55b63682dc248b842e9229f6db251c6e&chksm=871b0912b06c800437eedb3fab4ca179a8e83aac4c2879b218974a26e26f79cf6829aa6f4dfd&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌新论文：使用生成对抗网络的无监督像素级域适应  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720367&idx=3&sn=15551131dbd1d1fad15d2752dd78fe83&chksm=871b0c11b06c8507b17d846e7d09b8fc4f0f12267d2f7dd019afa3d0ca3345a8774041ef5c84&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌ICLR 2017论文提出超大规模的神经网络：稀疏门控专家混合层  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718466&idx=3&sn=ffeb503724abb0934c5c60e1202b12c0&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌论文：使用循环神经网络的全分辨率图像压缩  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722239&idx=4&sn=3544eee8b2bf5eeab1ae852ba9fde64c&chksm=871b0bc1b06c82d74fa57cbcb0d9f6178a78d474c6bfe93db174dd0c8f580272e398006b0f14&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌新论文提出适应性生成对抗网络AdaGAN：增强生成模型  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718943&idx=4&sn=f93b540a0b28100a6912e916d2ad1ac0&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌技术论文：用于油Tube推荐的深度神经网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719170&idx=3&sn=5c9dd730e1454335efdae14b24cd2053&chksm=871b07bcb06c8eaa7e4e210727f03214ddf633513522828ac2d3c5a339ee50486d5f32833566&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌与微软合著论文：由知识引导的结构化注意网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719594&idx=2&sn=b8398c3059b23babb02487baf2cc738f&chksm=871b0114b06c8802ed86e1bc0a17fb33d79500962206eb480bdf023b25c753a7535bb877df2b&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌深度解读：机器人可以如何通过共享经历学习新技能  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719099&idx=2&sn=52807674a2235e7ed8065a165427e1d6&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌NIPS 2016提交的8篇论文：从无监督学习到生成模型  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718429&idx=2&sn=2dad7f9aab23cf05323bb9d01c49d11b&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌DeepMind论文：使用合成梯度的解耦神经接口  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722118&idx=2&sn=42199ff21f3b06912e487de2c83eca1b&chksm=871b0b38b06c822ef121554a1f9bf23ef49e15cbbe6d84ac11c8e68eb377746a24809aa8e34c&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌提交ICLR 2017论文：学习记忆罕见事件  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650722472&idx=4&sn=892464db8e5e2d1e060bfeeacef63882&chksm=871b14d6b06c9dc0e5f870684b95dee11d47f75d155d601f86e9700ddc0c1b65191d9c46e19b&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌提出深度概率编程语言Edward：融合了贝叶斯、深度学习和概率编程 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721938&idx=4&sn=64634dd519ee70c35fdf1e771da7d9e5&chksm=871b0aecb06c83face511e46c836911a7a15c69980a76198ae928aa935b51d162ebe20946a35&scene=21#wechat_redirect" rel="nofollow,noindex">谷歌新论文提出预测器架构：端到端的学习与规划  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720220&idx=4&sn=85923b1102993f0c88c0394ae0ecb4fc&chksm=871b03a2b06c8ab4ca055cc73ee73c50275f70990ccdab71107626640548b916d383c549ea18&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind最新论文：线性时间的神经机器翻译  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718504&idx=3&sn=0469d037c1d004270da3006216a97cef&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind David Silver论文：学习跨多个数量级的值  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719494&idx=3&sn=67ed3e91731a47abe1a29df4c949119b&chksm=871b0178b06c886e6da4c83c0c26476c2f679f5e72d542531c5fbb8c798b6d768f9a9c596485&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind论文：在线Segment to Segment神经传导 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719745&idx=2&sn=242eda88fa9a5572e60b43e451a1549b&chksm=871b027fb06c8b693cff17f43553625f008fcb7965582119b651dbbd17e116e35dc801ebeec3&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind深度解读Nature论文：可微神经计算机  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719843&idx=4&sn=d54e7e415990bcfe69727e0b9f4c5f98&chksm=871b021db06c8b0b008c7004e130d596ae14f4b4de017098ecc4409e54fad8a4a77d3bcb93f4&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind论文：调控运动控制器的学习和迁移  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720969&idx=2&sn=d1fae404486906125e01b5def7e26d94&chksm=871b0eb7b06c87a1d3e7d0b29e8c8f9e4c86f4949d04b0485df1b45823555a6c88a25ccf4dc4&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind NIPS 2016论文盘点（Part1）：强化学习正大步向前  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721102&idx=2&sn=cbc44a149457d31ed9a8bbe825f09378&chksm=871b0f30b06c8626bb94cbd7a08fd4a5c8bec747082f49280fbd27e9c87c965de983a279796c&scene=21#wechat_redirect" rel="nofollow,noindex">DeepMind NIPS 2016论文盘点（Part2）：无监督学习的新进展  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719772&idx=2&sn=9540989a7862ef93d1e5146a7b5641c9&chksm=871b0262b06c8b74bcd961a7bfe45076a92c1d9456af235f8e84a4bce6e15d120f295cab5576&scene=21#wechat_redirect" rel="nofollow,noindex">ECCV 2016 最佳论文新鲜出炉  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719971&idx=3&sn=c23aa4b6172bd4ccea042d139ffa5daf&chksm=871b029db06c8b8b939ba2e0eb256db1d5fdd5ce2b229139cb31e7e54e5bdd2772761ce01c19&scene=21#wechat_redirect" rel="nofollow,noindex">Geoffrey Hinton论文：使用快速权重处理最近的过去 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717143&idx=2&sn=dbb912c06671f0ba4bb7faf8b1677831&scene=21#wechat_redirect" rel="nofollow,noindex">哈工大讯飞联合实验室最新论文刷新机器阅读理解纪录  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715759&idx=3&sn=20d16db66afb2d742422f0a3abfa7f1e&scene=21#wechat_redirect" rel="nofollow,noindex">华盛顿大学论文：使用机器学习分析科学文献中的视觉信息  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718786&idx=5&sn=1c91d80ce0422c7049bfdbaa8188d56f&scene=21#wechat_redirect" rel="nofollow,noindex">Ian Goodfellow 论文：通过视频预测的用于物理交互的无监督学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719909&idx=4&sn=90e189c8989817e3b52d2a40c355da10&chksm=871b02dbb06c8bcdc29a69b7a9ba2711f3ba1e2a08192e184f298093b404fa57272cdfdf81b9&scene=21#wechat_redirect" rel="nofollow,noindex">Ian Goodfellow 论文：用于隐私训练数据的深度学习的半监督知识迁移 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715900&idx=4&sn=948406fcbd53cdca4ce96bd33a28874d&scene=21#wechat_redirect" rel="nofollow,noindex">IBM论文：多尺度循环神经网络在对话生成中的应用  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715145&idx=3&sn=21d98866046180034d68da334501ce24&scene=21#wechat_redirect" rel="nofollow,noindex">ICLR2016会议，不可错过非死book提交的七篇论文  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718262&idx=1&sn=0391500e43530e7f4a8b8053176e5d7f&scene=21#wechat_redirect" rel="nofollow,noindex">计算机科学领导者：卡内基梅隆大学ACL2016论文汇总  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715794&idx=3&sn=3c328f66f24dee02b6a29a7821c7f6a1&scene=21#wechat_redirect" rel="nofollow,noindex">论文：TensorFlow，一个大规模机器学习系统  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717969&idx=4&sn=7fa67419573685604dde9811d434f6b8&scene=21#wechat_redirect" rel="nofollow,noindex">论文：通过连续奖励策略梯度学习在线比对  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718688&idx=4&sn=17b45f18ccb1e29e80fdca1645a71e5c&scene=21#wechat_redirect" rel="nofollow,noindex">论文：基准评测当前最先进的深度学习软件工具  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720189&idx=3&sn=8e0d8f7d3b9c037935736e09f85a4884&chksm=871b03c3b06c8ad57d13417933b779f6b0a5931c3448ef9bf2129fb64b0c3f4c84293358aebf&scene=21#wechat_redirect" rel="nofollow,noindex">论文：一种用于训练循环网络的新算法Professor Forcing </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718988&idx=4&sn=79c38974f2908e909e0f3bdd140cfa88&scene=21#wechat_redirect" rel="nofollow,noindex">论文：高斯混合模型的似然方法中的局部极大值：结构结果和算法结果 </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721081&idx=3&sn=111d844d50c98582695d04fa2b252c89&chksm=871b0f47b06c86514ac934c44fe4df92f5d4209f209b94b4c89d1f138492dbaf7e47479803a3&scene=21#wechat_redirect" rel="nofollow,noindex">NIPS 2016现场：谷歌发布 28 篇机器学习论文  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650715615&idx=1&sn=f114c4683656991bd331dc9971499a02&scene=21#wechat_redirect" rel="nofollow,noindex">Nature论文：无监督表征学习，用电子健康病历增强临床决策 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650716475&idx=1&sn=2b03deead0c1e63be80fdc239293e805&scene=21#wechat_redirect" rel="nofollow,noindex">Nature论文：从不确定性表征到自动建模  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718527&idx=2&sn=8d054db2eca7a0b25190a6cc050fc1d7&scene=21#wechat_redirect" rel="nofollow,noindex">NIPS 2016 公布571篇接收论文  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719581&idx=3&sn=e75b2fcd08d4ef5d8feeb79ebc223d18&chksm=871b0123b06c883511d8dbf853bb60f8e0903e335ed384421d1ab8db96ebf45d9b4cf3e96356&scene=21#wechat_redirect" rel="nofollow,noindex">OpenAI与NASA论文：用于张拉整体机器人运动的深度强化学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720257&idx=3&sn=4dfc480bfc70691f5baa44a75a0c1b82&chksm=871b0c7fb06c85696574d7f15d160afd3f45aa669bad4d1589f908c2fe0f56a3f34946bb6e73&scene=21#wechat_redirect" rel="nofollow,noindex">OpenAI论文：神经GPU的扩展和限制  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650721777&idx=1&sn=a51774b58c1b39ae9cb23c41361780af&chksm=871b098fb06c80992893818772a5c9410adb8826f60ae44d9468685413708c9c1c4583f967df&scene=21#wechat_redirect" rel="nofollow,noindex">苹果发布第一篇人工智能研究论文：模拟+无监督方法改善合成图像  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650716071&idx=2&sn=fb8944218199ac4c29af926d8847cc58&scene=21#wechat_redirect" rel="nofollow,noindex">商汤科技论文解析：人脸检测中级联卷积神经网络的联合训练 </a>  </p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717606&idx=3&sn=9d20357d6ce5877e7df31058c1ba0b4c&scene=21#wechat_redirect" rel="nofollow,noindex">斯坦福大学李飞飞最新论文：弱监督动作标记的连接时序模型  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720285&idx=5&sn=583751084a6855b35f684f582afd7976&chksm=871b0c63b06c857523341e94380cd6b87e84502854886f7aa8fa40ff665166db4ffda8a5ea1b&scene=21#wechat_redirect" rel="nofollow,noindex">Vicarious在ICLR2017提交无监督学习论文：层级组合特征学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719118&idx=3&sn=4565fb14db41208571b808956a39e310&scene=21#wechat_redirect" rel="nofollow,noindex">Yann LeCun论文：基于能量的生成对抗网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717728&idx=3&sn=2de2d33afc32c04f2bcb1cfe1a788c0f&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio 论文：一种神经知识语言模型  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720390&idx=1&sn=8ad603e853b88706ca4916495f59b228&chksm=871b0cf8b06c85ee216eb2d46361de137679cd9fd7d09a2a72cdf06d79d6c01dfe1683eb003c&scene=21#wechat_redirect" rel="nofollow,noindex">Yann LeCun提交ICLR 2017论文汇总：从GAN到循环实体网络等  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650717572&idx=2&sn=da7a53cde74285f229fcf6827bbb17d6&scene=21#wechat_redirect" rel="nofollow,noindex">Bengio论文：用于序列预测的actor-critic算法  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718133&idx=4&sn=97ce57340e916cb049ec96c96fa55fe3&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio论文：迈向生物学上可信的深度学习  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650719794&idx=4&sn=8a1d4e46663439c509f8d580e2b871e8&chksm=871b024cb06c8b5a1cbc79b9b2d9ff0d312842f04ad375ef7838c30f1b639cb3d1e1724f2902&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio论文：使用线性分类器探头理解中间层  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718391&idx=3&sn=75cac85169b49af157755a3cbc68670e&scene=21#wechat_redirect" rel="nofollow,noindex">Yoshua Bengio论文：Mollifying Networks </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650720220&idx=2&sn=0f2f5061e470ffbf76fee6ac77b3cf52&chksm=871b03a2b06c8ab4a0cb275b1044fb23a646997af91d5753439488ee976512c84979eadd2293&scene=21#wechat_redirect" rel="nofollow,noindex">微软重磅论文提出LightRNN：高效利用内存和计算的循环神经网络  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718106&idx=2&sn=93aceb9b6e4a0772bbaf9257a4def3d2&scene=21#wechat_redirect" rel="nofollow,noindex">微软ACL 2016论文汇集，自然语言技术逼近人类对话水平  </a></p>    <p><a href="http://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650718072&idx=2&sn=dc0f5e9ac4ca943afe91ea8d4e08f78c&scene=21#wechat_redirect" rel="nofollow,noindex">自然语言顶级会议ACL 2016谷歌论文汇集  </a></p>    <p> </p>    <p>来自：http://www.tuicool.com//articles/qQJbQjV</p>    <p> </p>