<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta charset="utf-8">
<meta name="generator" content="pdf2htmlEX">
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">
<link rel="stylesheet" href="https://static.pudn.com/base/css/base.min.css">
<link rel="stylesheet" href="https://static.pudn.com/base/css/fancy.min.css">
<link rel="stylesheet" href="https://static.pudn.com/prod/directory_preview_static/626c25111e41a87e8aaf119b/raw.css">
<script src="https://static.pudn.com/base/js/compatibility.min.js"></script>
<script src="https://static.pudn.com/base/js/pdf2htmlEX.min.js"></script>
<script>
try{
pdf2htmlEX.defaultViewer = new pdf2htmlEX.Viewer({});
}catch(e){}
</script>
<title></title>
</head>
<body>
<div id="sidebar" style="display: none">
<div id="outline">
</div>
</div>
<div id="pf1" class="pf w0 h0" data-page-no="1"><div class="pc pc1 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/626c25111e41a87e8aaf119b/bg1.jpg"><div class="c x0 y1 w2 h2"><div class="t m0 x1 h3 y2 ff1 fs0 fc0 sc0 ls0 ws0">第<span class="_ _0"> </span><span class="ff2 sc1">1<span class="_ _0"> </span></span>讲：深度学习基础和工具</div><div class="t m0 x2 h4 y3 ff1 fs1 fc0 sc0 ls0 ws0">——基本概念、<span class="_ _1"></span>基本模型、基本<span class="_ _1"></span>技术</div><div class="t m0 x3 h5 y4 ff1 fs2 fc0 sc0 ls0 ws0">东南大学数据与智能实验室 （<span class="_ _2"> </span><span class="ff3 sc1">D&Intel Lab<span class="_ _2"> </span></span>）</div><div class="t m0 x4 h6 y5 ff1 fs3 fc0 sc0 ls0 ws0">崇志宏</div><div class="t m0 x5 h6 y6 ff2 fs3 fc0 sc1 ls0 ws0">TEL<span class="_ _3"> </span><span class="ff1 sc0">：<span class="_ _3"> </span></span>13814066974</div><div class="t m0 x6 h7 y7 ff2 fs3 fc0 sc1 ls0 ws0">chongz<span class="_ _1"></span>hihong@<span class="_ _1"></span>seu.edu.<span class="_ _1"></span>cn</div><div class="t m0 x7 h7 y8 ff2 fs3 fc0 sc1 ls0 ws0">cse.se<span class="_ _1"></span>u.edu.c<span class="_ _1"></span>n/Person<span class="_ _1"></span>alPage/<span class="_ _1"></span>zhchong/</div><div class="t m0 x8 h7 y9 ff2 fs3 fc0 sc1 ls0 ws0">index.<span class="_ _1"></span>htm</div></div></div><div class="pi" data-data='{"ctm":[1.333333,0.000000,0.000000,1.333333,0.000000,0.000000]}'></div></div>
</body>
</html>
<div id="pf2" class="pf w0 h0" data-page-no="2"><div class="pc pc2 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/626c25111e41a87e8aaf119b/bg2.jpg"><div class="c x0 y1 w2 h2"><div class="t m0 x9 h8 ya ff1 fs4 fc1 sc2 ls0 ws0">深度学习是什么<span class="_ _4"> </span><span class="ff3 fc0 sc1">:<span class="ff4">①<span class="_ _4"> </span><span class="ff1 sc0">形式：万能的函数近似</span></span></span></div><div class="t m0 xa h5 y4 ff1 fs2 fc2 sc3 ls0 ws0"><span class="fc4 sc1">东南大学数据与智能实验室</span><span class="fc4 sc1"> </span><span class="_ _2"> </span><span class="ff3 sc1"><span class="fc4 sc1">D</span><span class="fc4 sc1">&In</span><span class="fc4 sc1">tel</span><span class="fc4 sc1"> </span><span class="fc4 sc1">L</span><span class="fc4 sc1">a</span><span class="fc4 sc1">b</span></span></div><div class="t m0 xb h9 yb ff1 fs5 fc3 sc4 ls0 ws0">深度神经网络<span class="_ _5"> </span><span class="ff3 sc1">=<span class="_ _5"> </span></span>高维复杂函数的近似</div><div class="t m0 xc h6 yc ff1 fs3 fc3 sc4 ls0 ws0"><span class="_ _3"> </span>相似算子激活算子合并算子</div></div></div><div class="pi" data-data='{"ctm":[1.333333,0.000000,0.000000,1.333333,0.000000,0.000000]}'></div></div>
<div id="pf3" class="pf w0 h0" data-page-no="3"><div class="pc pc3 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/626c25111e41a87e8aaf119b/bg3.jpg"><div class="c x0 y1 w2 h2"><div class="t m0 xa h5 y4 ff1 fs2 fc2 sc3 ls0 ws0"><span class="fc4 sc1">东南大学数据与智能实验室</span><span class="fc4 sc1"> </span><span class="_ _2"> </span><span class="ff3 sc1"><span class="fc4 sc1">D</span><span class="fc4 sc1">&In</span><span class="fc4 sc1">tel</span><span class="fc4 sc1"> </span><span class="fc4 sc1">L</span><span class="fc4 sc1">a</span><span class="fc4 sc1">b</span></span></div><div class="t m0 xd h6 yd ff1 fs3 fc3 sc4 ls0 ws0">学习<span class="_ _3"> </span><span class="ff3 sc1">=><span class="_ _6"> </span></span>优化</div><div class="t m0 x9 h8 ye ff1 fs4 fc1 sc2 ls0 ws0">深度学习是什么<span class="_ _4"> </span><span class="ff3 fc0 sc1">:<span class="ff4">②<span class="_ _4"> </span><span class="ff1 sc0">学习：目标、优化、泛</span></span></span></div><div class="t m0 x9 h8 yf ff1 fs4 fc0 sc0 ls0 ws0">化</div></div></div><div class="pi" data-data='{"ctm":[1.333333,0.000000,0.000000,1.333333,0.000000,0.000000]}'></div></div>
<div id="pf4" class="pf w0 h0" data-page-no="4"><div class="pc pc4 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/626c25111e41a87e8aaf119b/bg4.jpg"><div class="c x0 y1 w2 h2"><div class="t m0 x3 h5 y4 ff1 fs2 fc0 sc0 ls0 ws0">东南大学数据与智能实验室（ <span class="_ _2"> </span><span class="ff3 sc1">D&Intel Lab<span class="_ _2"> </span></span>）</div><div class="t m0 x9 h8 ya ff1 fs4 fc1 sc2 ls0 ws0">深度学习是什么<span class="fc0 sc0">：例如</span></div></div></div><div class="pi" data-data='{"ctm":[1.333333,0.000000,0.000000,1.333333,0.000000,0.000000]}'></div></div>
<div id="pf5" class="pf w0 h0" data-page-no="5"><div class="pc pc5 w0 h0"><img class="bi x0 y0 w1 h1" alt="" src="https://static.pudn.com/prod/directory_preview_static/626c25111e41a87e8aaf119b/bg5.jpg"><div class="c x0 y1 w2 h2"><div class="t m0 xe ha y10 ff2 fs6 fc0 sc1 ls0 ws0">I. <span class="_ _7"> </span><span class="ff1 sc0">机器学习<span class="_ _7"> </span></span>/<span class="_ _7"> </span><span class="ff1 sc0">深度学习基础（第一部分<span class="_ _8"></span>）</span></div><div class="t m0 xf ha y11 ff3 fs6 fc0 sc1 ls0 ws0">1<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">基本概念、方法和<span class="_ _7"> </span></span><span class="ff2">BP<span class="_ _7"> </span><span class="ff1 sc0">算法：计算效率和统<span class="_ _8"></span>计效率</span></span></div><div class="t m0 xf ha y12 ff3 fs6 fc0 sc1 ls0 ws0">2<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">大规模深度学习工具和<span class="_ _8"></span>实践</span></div><div class="t m0 xe ha y13 ff2 fs6 fc0 sc1 ls0 ws0">II. <span class="_ _7"> </span><span class="ff1 sc0">基本技术、<span class="_ _8"></span>过程和基本模型<span class="_ _8"></span>以及应用案例（第二部分<span class="_ _8"></span>）</span></div><div class="t m0 xf ha y14 ff3 fs6 fc0 sc1 ls0 ws0">3<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">深度学习基本技术、过<span class="_ _8"></span>程</span></div><div class="t m0 xf ha y15 ff3 fs6 fc0 sc1 ls0 ws0">4<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">卷积神经网络<span class="_ _7"> </span></span><span class="ff2">/<span class="_ _7"> </span><span class="ff1 sc0">递归神经网络（<span class="_ _7"> </span></span>LSTM<span class="_ _7"> </span><span class="ff1 sc0">、<span class="_ _7"> </span></span>GRU<span class="_ _7"> </span><span class="ff1 sc0">、<span class="_ _7"> </span></span>Bi-RNN </span></div><div class="t m0 x10 ha y16 ff2 fs6 fc0 sc1 ls0 ws0">Atten<span class="_ _8"></span>tion<span class="_ _8"></span> base<span class="_ _8"></span>d RNN<span class="_ _3"> </span><span class="ff1 sc0">） </span></div><div class="t m0 x10 ha y17 ff2 fs6 fc0 sc1 ls0 ws0">/<span class="_ _7"> </span><span class="ff1 sc0">残差网络<span class="_ _7"> </span></span>/<span class="_ _7"> </span><span class="ff1 sc0">平移、扭曲不变性<span class="_ _7"> </span></span>/Seq2Seq<span class="_ _b"> </span><span class="ff1 sc0">的原理与实现</span></div><div class="t m0 xf ha y18 ff2 fs6 fc0 sc1 ls0 ws0">5.<span class="_"> </span>R/Faster<span class="_ _8"></span>/Mas<span class="_ _8"></span>k-CNN<span class="_ _8"></span> <span class="_ _7"> </span><span class="ff1 sc0">（<span class="_ _7"> </span></span>SPPNE<span class="_ _8"></span>T<span class="_ _7"> </span><span class="ff1 sc0">）、<span class="_ _7"> </span></span>PixelNe<span class="_ _8"></span>t<span class="_ _7"> </span><span class="ff1 sc0">、<span class="_ _7"> </span></span>YOLO<span class="_ _7"> </span><span class="ff1 sc0">、<span class="_ _7"> </span></span>SSD</div><div class="t m0 xe ha y19 ff2 fs6 fc0 sc1 ls0 ws0">III. <span class="_ _7"> </span><span class="ff1 sc0">基本<span class="_ _8"></span>学习范式（第三部分）</span></div><div class="t m0 xf ha y1a ff3 fs6 fc0 sc1 ls0 ws0">6<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">学习目标和关键技术：<span class="_ _8"></span>似然和非似然估计<span class="_ _8"></span>（<span class="_ _c"> </span></span><span class="ff2">VAE<span class="_ _7"> </span><span class="ff1 sc0">和<span class="_ _c"> </span></span>G<span class="_ _8"></span>AN<span class="_ _7"> </span><span class="ff1 sc0">、<span class="_ _c"> </span></span>WGA<span class="_ _8"></span>N<span class="_ _7"> </span><span class="ff1 sc0">）、</span></span></div><div class="t m0 x10 ha y1b ff1 fs6 fc0 sc0 ls0 ws0">指数<span class="_ _c"> </span><span class="ff2 sc1">/<span class="_ _7"> </span></span>下界和后验<span class="_ _8"></span>推理<span class="_ _c"> </span><span class="ff2 sc1">/<span class="_ _7"> </span></span>重参数</div><div class="t m0 xf ha y1c ff3 fs6 fc0 sc1 ls0 ws0">7<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">无导师学习：自编码框<span class="_ _8"></span>架<span class="_ _c"> </span></span><span class="ff2">/<span class="_ _7"> </span><span class="ff1 sc0">生成对抗<span class="_ _7"> </span></span>/<span class="_ _c"> </span><span class="ff1 sc0">博弈学习<span class="_ _7"> </span></span>/<span class="_ _7"> </span><span class="ff1 sc0">对偶学习</span></span></div><div class="t m0 xf ha y1d ff3 fs6 fc0 sc1 ls0 ws0">8<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">弱监督学习<span class="_ _7"> </span></span><span class="ff2">/<span class="_ _c"> </span><span class="ff1 sc0">转移学习</span></span></div><div class="t m0 xf ha y1e ff3 fs6 fc0 sc1 ls0 ws0">9<span class="_ _9"></span>.<span class="_ _a"> </span><span class="ff1 sc0">强化学习和深度强化学<span class="_ _8"></span>习</span></div><div class="t m0 xe ha y1f ff2 fs6 fc0 sc1 ls0 ws0">IV. <span class="_ _7"> </span><span class="ff1 sc0">深度学习设计<span class="_ _8"></span>模式和模型调试（第<span class="_ _8"></span>四部分）</span></div><div class="t m0 xf ha y20 ff3 fs6 fc0 sc1 ls0 ws0">1<span class="_ _9"></span>0<span class="_ _9"></span>.<span class="_ _2"> </span><span class="ff1 sc0">记忆组件和关注<span class="_ _8"></span>技术</span></div><div class="t m0 xf ha y21 ff3 fs6 fc0 sc1 ls0 ws0">1<span class="_ _9"></span>1<span class="_ _9"></span>.<span class="_ _2"> </span><span class="ff1 sc0">多模态数据和深<span class="_ _8"></span>度社交推荐</span></div><div class="t m0 xf ha y22 ff3 fs6 fc0 sc1 ls0 ws0">1<span class="_ _9"></span>2<span class="_ _9"></span>.<span class="_ _2"> </span><span class="ff1 sc0">贝叶斯深度学习<span class="_ _7"> </span></span><span class="ff2">/<span class="_ _7"> </span><span class="ff1 sc0">深度<span class="_ _c"> </span></span>LD<span class="_ _8"></span>A/<span class="_ _7"> </span><span class="ff1 sc0">组件组合学习</span></span></div><div class="t m0 xf ha y23 ff3 fs6 fc0 sc1 ls0 ws0">1<span class="_ _9"></span>3<span class="_ _9"></span>.<span class="_ _2"> </span><span class="ff1 sc0">模型设计和模型<span class="_ _8"></span>调试（训练目标设<span class="_ _8"></span>计、不稳定、超参数设<span class="_ _8"></span>计自动化）</span></div><div class="t m0 x11 hb y24 ff1 fs7 fc1 sc2 ls0 ws0">计算效率和统<span class="_ _1"></span>计效率</div><div class="t m0 x11 hb y25 ff2 fs7 fc3 sc1 ls0 ws0">1<span class="_ _b"> </span><span class="ff1 sc4">、基本概念</span></div><div class="t m0 x11 hb y26 ff2 fs7 fc3 sc1 ls0 ws0">2<span class="_ _b"> </span><span class="ff1 sc4">、基本模型</span></div><div class="t m0 x11 hb y27 ff2 fs7 fc3 sc1 ls0 ws0">3<span class="_ _b"> </span><span class="ff1 sc4">、基本技术</span></div><div class="t m0 x11 hb y28 ff2 fs7 fc3 sc1 ls0 ws0">4<span class="_ _b"> </span><span class="ff1 sc4">、模型设计模<span class="_ _1"></span>式和调试</span></div></div></div><div class="pi" data-data='{"ctm":[1.333333,0.000000,0.000000,1.333333,0.000000,0.000000]}'></div></div>
