-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
371 lines (331 loc) · 28.3 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
<!DOCTYPE HTML>
<html lang="en" class="sidebar-visible no-js light">
<head>
<!-- Book generated using mdBook -->
<meta charset="UTF-8">
<title>引言 - Machine Learning Yearning</title>
<!-- Custom HTML head -->
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
<meta name="description" content="">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="theme-color" content="#ffffff" />
<link rel="icon" href="favicon.svg">
<link rel="shortcut icon" href="favicon.png">
<link rel="stylesheet" href="css/variables.css">
<link rel="stylesheet" href="css/general.css">
<link rel="stylesheet" href="css/chrome.css">
<link rel="stylesheet" href="css/print.css" media="print">
<!-- Fonts -->
<link rel="stylesheet" href="FontAwesome/css/font-awesome.css">
<link rel="stylesheet" href="fonts/fonts.css">
<!-- Highlight.js Stylesheets -->
<link rel="stylesheet" href="highlight.css">
<link rel="stylesheet" href="tomorrow-night.css">
<link rel="stylesheet" href="ayu-highlight.css">
<!-- Custom theme stylesheets -->
</head>
<body>
<!-- Provide site root to javascript -->
<script type="text/javascript">
var path_to_root = "";
var default_theme = window.matchMedia("(prefers-color-scheme: dark)").matches ? "navy" : "light";
</script>
<!-- Work around some values being stored in localStorage wrapped in quotes -->
<script type="text/javascript">
try {
var theme = localStorage.getItem('mdbook-theme');
var sidebar = localStorage.getItem('mdbook-sidebar');
if (theme.startsWith('"') && theme.endsWith('"')) {
localStorage.setItem('mdbook-theme', theme.slice(1, theme.length - 1));
}
if (sidebar.startsWith('"') && sidebar.endsWith('"')) {
localStorage.setItem('mdbook-sidebar', sidebar.slice(1, sidebar.length - 1));
}
} catch (e) { }
</script>
<!-- Set the theme before any content is loaded, prevents flash -->
<script type="text/javascript">
var theme;
try { theme = localStorage.getItem('mdbook-theme'); } catch(e) { }
if (theme === null || theme === undefined) { theme = default_theme; }
var html = document.querySelector('html');
html.classList.remove('no-js')
html.classList.remove('light')
html.classList.add(theme);
html.classList.add('js');
</script>
<!-- Hide / unhide sidebar before it is displayed -->
<script type="text/javascript">
var html = document.querySelector('html');
var sidebar = 'hidden';
if (document.body.clientWidth >= 1080) {
try { sidebar = localStorage.getItem('mdbook-sidebar'); } catch(e) { }
sidebar = sidebar || 'visible';
}
html.classList.remove('sidebar-visible');
html.classList.add("sidebar-" + sidebar);
</script>
<nav id="sidebar" class="sidebar" aria-label="Table of contents">
<div class="sidebar-scrollbox">
<ol class="chapter"><li class="chapter-item expanded affix "><a href="index.html" class="active">引言</a></li><li class="chapter-item expanded "><a href="chapter1.html"><strong aria-hidden="true">1.</strong> 机器学习策略的原因</a></li><li class="chapter-item expanded "><a href="chapter2.html"><strong aria-hidden="true">2.</strong> 如何使用本书来帮助您的团队</a></li><li class="chapter-item expanded "><a href="chapter3.html"><strong aria-hidden="true">3.</strong> 预备知识和注释</a></li><li class="chapter-item expanded "><a href="chapter4.html"><strong aria-hidden="true">4.</strong> 规模推动机器学习进步</a></li><li class="chapter-item expanded "><a href="chapter5.html"><strong aria-hidden="true">5.</strong> 您的开发和测试集</a></li><li class="chapter-item expanded "><a href="chapter6.html"><strong aria-hidden="true">6.</strong> 你的开发集和测试集应该来自相同的分布</a></li><li class="chapter-item expanded "><a href="chapter7.html"><strong aria-hidden="true">7.</strong> 开发集/测试集需要多大</a></li><li class="chapter-item expanded "><a href="chapter8.html"><strong aria-hidden="true">8.</strong> 为您的团队建立单一数字的评估指标以进行优化</a></li><li class="chapter-item expanded "><a href="chapter9.html"><strong aria-hidden="true">9.</strong> 优化指标和满足指标</a></li><li class="chapter-item expanded "><a href="chapter10.html"><strong aria-hidden="true">10.</strong> 通过开发集和评估标准加速迭代</a></li><li class="chapter-item expanded "><a href="chapter11.html"><strong aria-hidden="true">11.</strong> 何时更改开发/测试集和评估指标</a></li><li class="chapter-item expanded "><a href="chapter12.html"><strong aria-hidden="true">12.</strong> 小结：建立开发集和测试集</a></li><li class="chapter-item expanded "><a href="chapter13.html"><strong aria-hidden="true">13.</strong> 快速构建您的第一个系统，然后迭代</a></li><li class="chapter-item expanded "><a href="chapter14.html"><strong aria-hidden="true">14.</strong> 误差分析：查看开发集样本以评估想法</a></li><li class="chapter-item expanded "><a href="chapter15.html"><strong aria-hidden="true">15.</strong> 在误差分析期间并行评估多个想法</a></li><li class="chapter-item expanded "><a href="chapter16.html"><strong aria-hidden="true">16.</strong> 清理错误标注的开发和测试集样本</a></li><li class="chapter-item expanded "><a href="chapter17.html"><strong aria-hidden="true">17.</strong> 如果你有一个大的开发集，将其分成两个子集，只着眼于其中的一个</a></li><li class="chapter-item expanded "><a href="chapter18.html"><strong aria-hidden="true">18.</strong> Eyeball 和 Blackbox 开发集应该多大？</a></li><li class="chapter-item expanded "><a href="chapter19.html"><strong aria-hidden="true">19.</strong> 小贴士：基本误差分析</a></li><li class="chapter-item expanded "><a href="chapter20.html"><strong aria-hidden="true">20.</strong> 偏差和方差：误差的两大来源</a></li><li class="chapter-item expanded "><a href="chapter21.html"><strong aria-hidden="true">21.</strong> 偏差和方差的例子</a></li><li class="chapter-item expanded "><a href="chapter22.html"><strong aria-hidden="true">22.</strong> 比较最优错误率</a></li><li class="chapter-item expanded "><a href="chapter23.html"><strong aria-hidden="true">23.</strong> 处理偏差和方差</a></li><li class="chapter-item expanded "><a href="chapter24.html"><strong aria-hidden="true">24.</strong> 偏差和方差间的权衡</a></li><li class="chapter-item expanded "><a href="chapter25.html"><strong aria-hidden="true">25.</strong> 减少可避免偏差的方法</a></li><li class="chapter-item expanded "><a href="chapter26.html"><strong aria-hidden="true">26.</strong> 训练集上的误差分析</a></li><li class="chapter-item expanded "><a href="chapter27.html"><strong aria-hidden="true">27.</strong> 减少方差的方法</a></li><li class="chapter-item expanded "><a href="chapter28.html"><strong aria-hidden="true">28.</strong> 诊断偏差和方差：学习曲线</a></li><li class="chapter-item expanded "><a href="chapter29.html"><strong aria-hidden="true">29.</strong> 绘制训练误差曲线</a></li><li class="chapter-item expanded "><a href="chapter30.html"><strong aria-hidden="true">30.</strong> 解读学习曲线：高偏差</a></li><li class="chapter-item expanded "><a href="chapter31.html"><strong aria-hidden="true">31.</strong> 解释学习曲线：其他情况</a></li><li class="chapter-item expanded "><a href="chapter32.html"><strong aria-hidden="true">32.</strong> 绘制学习曲线</a></li><li class="chapter-item expanded "><a href="chapter33.html"><strong aria-hidden="true">33.</strong> 为何我们要与人类水平的表现作对比</a></li><li class="chapter-item expanded "><a href="chapter34.html"><strong aria-hidden="true">34.</strong> 如何定义人类水平的表现</a></li><li class="chapter-item expanded "><a href="chapter35.html"><strong aria-hidden="true">35.</strong> 超越人类水平表现</a></li><li class="chapter-item expanded "><a href="chapter36.html"><strong aria-hidden="true">36.</strong> 何时应该在不同的分布下训练和测试</a></li><li class="chapter-item expanded "><a href="chapter37.html"><strong aria-hidden="true">37.</strong> 如何决定是否使用所有数据</a></li><li class="chapter-item expanded "><a href="chapter38.html"><strong aria-hidden="true">38.</strong> 如何决定是否包含不一致的数据</a></li><li class="chapter-item expanded "><a href="chapter39.html"><strong aria-hidden="true">39.</strong> 加权数据</a></li><li class="chapter-item expanded "><a href="chapter40.html"><strong aria-hidden="true">40.</strong> 从训练集到开发集的泛化</a></li><li class="chapter-item expanded "><a href="chapter41.html"><strong aria-hidden="true">41.</strong> 识别偏差、方差和数据不匹配误差</a></li><li class="chapter-item expanded "><a href="chapter42.html"><strong aria-hidden="true">42.</strong> 处理数据不匹配</a></li><li class="chapter-item expanded "><a href="chapter43.html"><strong aria-hidden="true">43.</strong> 人工数据合成</a></li><li class="chapter-item expanded "><a href="chapter44.html"><strong aria-hidden="true">44.</strong> 优化验证测试</a></li><li class="chapter-item expanded "><a href="chapter45.html"><strong aria-hidden="true">45.</strong> 优化验证集的一般形式</a></li><li class="chapter-item expanded "><a href="chapter46.html"><strong aria-hidden="true">46.</strong> 强化学习样本</a></li><li class="chapter-item expanded "><a href="chapter47.html"><strong aria-hidden="true">47.</strong> 端到端学习的兴起</a></li><li class="chapter-item expanded "><a href="chapter48.html"><strong aria-hidden="true">48.</strong> 更多端到端学习示例</a></li><li class="chapter-item expanded "><a href="chapter49.html"><strong aria-hidden="true">49.</strong> 端到端学习的优点和缺点</a></li><li class="chapter-item expanded "><a href="chapter50.html"><strong aria-hidden="true">50.</strong> 选择流水线组件：数据可用性</a></li><li class="chapter-item expanded "><a href="chapter51.html"><strong aria-hidden="true">51.</strong> 选择流水线组件：任务简单</a></li><li class="chapter-item expanded "><a href="chapter52.html"><strong aria-hidden="true">52.</strong> 直接学习丰富的输出</a></li><li class="chapter-item expanded "><a href="chapter53.html"><strong aria-hidden="true">53.</strong> 组件错误分析</a></li><li class="chapter-item expanded "><a href="chapter54.html"><strong aria-hidden="true">54.</strong> 将错误归因于某个组件</a></li><li class="chapter-item expanded "><a href="chapter55.html"><strong aria-hidden="true">55.</strong> 错误归因的一般情况</a></li><li class="chapter-item expanded "><a href="chapter56.html"><strong aria-hidden="true">56.</strong> 组件错误分析和与人类水平的对比</a></li><li class="chapter-item expanded "><a href="chapter57.html"><strong aria-hidden="true">57.</strong> 发现有瑕疵的ML流水线</a></li><li class="chapter-item expanded "><a href="chapter58.html"><strong aria-hidden="true">58.</strong> 组建一个超级英雄团队——让你的队友阅读本书</a></li></ol>
</div>
<div id="sidebar-resize-handle" class="sidebar-resize-handle"></div>
</nav>
<div id="page-wrapper" class="page-wrapper">
<div class="page">
<div id="menu-bar-hover-placeholder"></div>
<div id="menu-bar" class="menu-bar sticky bordered">
<div class="left-buttons">
<button id="sidebar-toggle" class="icon-button" type="button" title="Toggle Table of Contents" aria-label="Toggle Table of Contents" aria-controls="sidebar">
<i class="fa fa-bars"></i>
</button>
<button id="theme-toggle" class="icon-button" type="button" title="Change theme" aria-label="Change theme" aria-haspopup="true" aria-expanded="false" aria-controls="theme-list">
<i class="fa fa-paint-brush"></i>
</button>
<ul id="theme-list" class="theme-popup" aria-label="Themes" role="menu">
<li role="none"><button role="menuitem" class="theme" id="light">Light (default)</button></li>
<li role="none"><button role="menuitem" class="theme" id="rust">Rust</button></li>
<li role="none"><button role="menuitem" class="theme" id="coal">Coal</button></li>
<li role="none"><button role="menuitem" class="theme" id="navy">Navy</button></li>
<li role="none"><button role="menuitem" class="theme" id="ayu">Ayu</button></li>
</ul>
<button id="search-toggle" class="icon-button" type="button" title="Search. (Shortkey: s)" aria-label="Toggle Searchbar" aria-expanded="false" aria-keyshortcuts="S" aria-controls="searchbar">
<i class="fa fa-search"></i>
</button>
</div>
<h1 class="menu-title">Machine Learning Yearning</h1>
<div class="right-buttons">
<a href="print.html" title="Print this book" aria-label="Print this book">
<i id="print-button" class="fa fa-print"></i>
</a>
</div>
</div>
<div id="search-wrapper" class="hidden">
<form id="searchbar-outer" class="searchbar-outer">
<input type="search" id="searchbar" name="searchbar" placeholder="Search this book ..." aria-controls="searchresults-outer" aria-describedby="searchresults-header">
</form>
<div id="searchresults-outer" class="searchresults-outer hidden">
<div id="searchresults-header" class="searchresults-header"></div>
<ul id="searchresults">
</ul>
</div>
</div>
<!-- Apply ARIA attributes after the sidebar and the sidebar toggle button are added to the DOM -->
<script type="text/javascript">
document.getElementById('sidebar-toggle').setAttribute('aria-expanded', sidebar === 'visible');
document.getElementById('sidebar').setAttribute('aria-hidden', sidebar !== 'visible');
Array.from(document.querySelectorAll('#sidebar a')).forEach(function(link) {
link.setAttribute('tabIndex', sidebar === 'visible' ? 0 : -1);
});
</script>
<div id="content" class="content">
<main>
<h1 id="machine-learning-yearning"><a class="header" href="#machine-learning-yearning">Machine Learning Yearning</a></h1>
<p><strong>目 录</strong></p>
<ul>
<li><a href="#machine-learning-yearning">Machine Learning Yearning</a>
<ul>
<li><a href="#%E7%AE%80%E4%BB%8B">简介</a></li>
<li><a href="#%E7%9B%AE%E7%9A%84">目的</a></li>
<li><a href="#%E7%BF%BB%E8%AF%91%E7%AB%A0%E8%8A%82">翻译章节</a>
<ul>
<li><a href="#setting-up-development-and-test-sets">Setting up development and test sets</a></li>
<li><a href="#basic-error-analysis">Basic Error Analysis</a></li>
<li><a href="#bias-and-variance">Bias and Variance</a></li>
<li><a href="#learning-curves">Learning curves</a></li>
<li><a href="#comparing-to-human-level-performance">Comparing to human-level performance</a></li>
<li><a href="#training-and-testing-on-different-distributions">Training and testing on different distributions</a></li>
<li><a href="#debugging-inference-algorithms">Debugging inference algorithms</a></li>
<li><a href="#end-to-end-deep-learning">End-to-end deep learning</a></li>
<li><a href="#error-analysis-by-parts">Error analysis by parts</a></li>
<li><a href="#conclusion">Conclusion</a></li>
</ul>
</li>
<li><a href="#%E8%8B%B1%E6%96%87%E5%8E%9F%E6%96%87">英文原文</a></li>
</ul>
</li>
</ul>
<h2 id="简介"><a class="header" href="#简介">简介</a></h2>
<p>NG的手稿，已出全。我这里边学习边翻译，随手记录之，加深学习印象，仅供学习交流。 </p>
<p>官网：<a href="http://www.mlyearning.org/">http://www.mlyearning.org/</a></p>
<p>更好阅读体验，移步gitbook：https://xiaqunfeng.gitbooks.io/machine-learning-yearning/content/</p>
<p>epub 格式下载： <a href="draft/Machine%2520Learning%2520Yearning.epub">Machine Learning Yearning.epub</a></p>
<blockquote>
<p><strong>声明</strong>：本rep是自己学习过程的一个记录，仅用于学习目的。</p>
</blockquote>
<p><strong>更新记录：</strong></p>
<ul>
<li>update 2018.04.25：NG终于出15~19章的手稿啦，等的好辛苦（DONE）</li>
</ul>
<blockquote>
<p>Tips：在原先的12章和13章之间新增一个章节 <code>13 Build your first system quickly, then iterate</code>，原先的chapter13变为14，chapter14变为15</p>
</blockquote>
<ul>
<li>update 2018.05.02：手稿 20~22 章已出（DONE）</li>
<li>update 2018.05.09：手稿 23~27 章已出（DONE）</li>
<li>update 2018.05.16：手稿 28~30 章已出（DONE）</li>
<li>update 2018.05.23：手稿 31~32 章已出（DONE）</li>
<li>update 2018.05.30：手稿 33~35 章已出（DONE）</li>
<li>update 2018.06.06：手稿 36~39 章已出（DONE）</li>
<li>update 2018.06.13：手稿 40~43 章已出（DONE）</li>
<li>update 2018.06.20：手稿 44~46 章已出（DONE）</li>
<li>update 2018.06.27：手稿 47~49 章已出（DONE）</li>
<li>update 2018.07.04：手稿 50~52 章已出（DONE）</li>
<li>update 2018.09.29：手稿 53~58 章已出（DOING）</li>
</ul>
<blockquote>
<p>业余时间翻译，水平有限，如有不妥或错误之处，欢迎不吝赐教。</p>
</blockquote>
<h2 id="目的"><a class="header" href="#目的">目的</a></h2>
<p>这本书的目的是教你如何做组织一个机器学习项目所需的大量的决定。 你将学习：</p>
<ul>
<li>
<p>如何建立你的开发和测试集</p>
</li>
<li>
<p>基本错误分析</p>
</li>
<li>
<p>如何使用偏差和方差来决定该做什么</p>
</li>
<li>
<p>学习曲线</p>
</li>
<li>
<p>将学习算法与人类水平的表现进行比较</p>
</li>
<li>
<p>调试推理算法</p>
</li>
<li>
<p>什么时候应该和不应该使用端到端的深度学习</p>
</li>
<li>
<p>按步进行错误分析</p>
</li>
</ul>
<h2 id="翻译章节"><a class="header" href="#翻译章节">翻译章节</a></h2>
<p><a href="chapter1.html">Chapter 1、Why Machine Learning Strategy</a></p>
<p><a href="chapter2.html">Chapter 2、How to use this book to help your team</a></p>
<p><a href="chapter3.html">Chapter 3、Prerequisites and Notation</a></p>
<p><a href="chapter4.html">Chapter 4、Scale drives machine learning progress</a></p>
<h3 id="setting-up-development-and-test-sets"><a class="header" href="#setting-up-development-and-test-sets">Setting up development and test sets</a></h3>
<p><a href="chapter5.html">Chapter 5、Your development and test sets</a></p>
<p><a href="chapter6.html">Chapter 6、Your dev and test sets should come from the same distribution</a></p>
<p><a href="chapter7.html">Chapter 7、How large do the dev/test sets need to be?</a></p>
<p><a href="chapter8.html">Chapter 8、Establish a single-number evaluation metric for your team to optimize</a></p>
<p><a href="chapter9.html">Chapter 9、Optimizingandsatisficingmetrics</a></p>
<p><a href="chapter10.html">Chapter 10、Having a dev set and metric speeds up iterations</a></p>
<p><a href="chapter11.html">Chapter 11、When to change dev/test sets and metrics</a></p>
<p><a href="chapter12.html">Chapter 12、Takeaways: Setting up development and test sets</a></p>
<h3 id="basic-error-analysis"><a class="header" href="#basic-error-analysis">Basic Error Analysis</a></h3>
<p><a href="chapter13.html">Chapter 13、Build your first system quickly, then iterate</a></p>
<p><a href="chapter14.html">Chapter 14、Error analysis: Look at dev set examples to evaluate ideas</a></p>
<p><a href="chapter15.html">Chapter 15、Evaluate multiple ideas in parallel during error analysis</a></p>
<p><a href="chapter16.html">Chapter 16、Cleaning up mislabeled dev and test set examples</a></p>
<p><a href="chapter17.html">Chapter 17、 If you have a large dev set, split it into two subsets, only one of which you look at</a></p>
<p><a href="chapter18.html">Chapter 18、How big should the Eyeball and Blackbox dev sets be?</a></p>
<p><a href="chapter19.html">Chapter 19、Takeaways: Basic error analysis</a></p>
<h3 id="bias-and-variance"><a class="header" href="#bias-and-variance">Bias and Variance</a></h3>
<p><a href="chapter20.html">Chapter 20、Bias and Variance: The two big sources of error</a></p>
<p><a href="chapter21.html">Chapter 21、Examples of Bias and Variance</a></p>
<p><a href="chapter22.html">Chapter 22、Comparing to the optimal error rate</a></p>
<p><a href="chapter23.html">Chapter 23、Addressing Bias and Variance</a></p>
<p><a href="chapter24.html">Chapter 24、Bias vs. Variance tradeoff</a></p>
<p><a href="chapter25.html">Chapter 25、Techniques for reducing avoidable bias</a></p>
<p><a href="chapter26.html">Chapter 26、Error analysis on the training set</a></p>
<p><a href="chapter27.html">Chapter 27、Techniques for reducing variance</a></p>
<h3 id="learning-curves"><a class="header" href="#learning-curves">Learning curves</a></h3>
<p><a href="chapter28.html">Chapter 28、Diagnosing bias and variance: Learning curves</a></p>
<p><a href="chapter29.html">Chapter 29、Plotting training error</a></p>
<p><a href="chapter30.html">Chapter 30、Interpreting learning curves: High bias</a></p>
<p><a href="chapter31.html">Chapter 31、Interpreting learning curves: Other cases</a></p>
<p><a href="chapter32.html">Chapter 32、Plotting learning curves</a></p>
<h3 id="comparing-to-human-level-performance"><a class="header" href="#comparing-to-human-level-performance">Comparing to human-level performance</a></h3>
<p><a href="chapter33.html">Chapter 33、Why we compare to human-level performance</a></p>
<p><a href="chapter34.html">Chapter 34、How to define human-level performance</a></p>
<p><a href="chapter35.html">Chapter 35、Surpassing human-level performance</a></p>
<h3 id="training-and-testing-on-different-distributions"><a class="header" href="#training-and-testing-on-different-distributions">Training and testing on different distributions</a></h3>
<p><a href="chapter36.html">Chapter 36、When you should train and test on different distributions</a></p>
<p><a href="chapter37.html">Chapter 37、How to decide whether to use all your data</a></p>
<p><a href="chapter38.html">Chapter 38、How to decide whether to include inconsistent data</a></p>
<p><a href="chapter39.html">Chapter 39、Weighting data</a></p>
<p><a href="chapter40.html">Chapter 40、Generalizing from the training set to the dev set</a></p>
<p><a href="chapter41.html">Chapter 41、Identifying Bias, Variance, and Data Mismatch Errors</a></p>
<p><a href="chapter42.html">Chapter 42、Addressing data mismatch</a></p>
<p><a href="chapter43.html">Chapter 43、Artificial data synthesis</a></p>
<h3 id="debugging-inference-algorithms"><a class="header" href="#debugging-inference-algorithms">Debugging inference algorithms</a></h3>
<p><a href="chapter44.html">Chapter 44、The Optimization Verification test</a></p>
<p><a href="chapter45.html">Chapter 45、General form of Optimization Verification test</a></p>
<p><a href="chapter46.html">Chapter 46、Reinforcement learning example</a></p>
<h3 id="end-to-end-deep-learning"><a class="header" href="#end-to-end-deep-learning">End-to-end deep learning</a></h3>
<p><a href="chapter47.html">Chapter 47、The rise of end-to-end learning</a></p>
<p><a href="chapter48.html">Chapter 48、More end-to-end learning examples</a></p>
<p><a href="chapter49.html">Chapter 49、Pros and cons of end-to-end learning</a></p>
<p><a href="chapter50.html">Chapter 50、Choosing pipeline components: Data availability</a></p>
<p><a href="chapter51.html">Chapter 51、Choosing pipeline components: Task simplicity</a></p>
<p><a href="chapter52.html">Chapter 52、Directly learning rich outputs</a></p>
<h3 id="error-analysis-by-parts"><a class="header" href="#error-analysis-by-parts">Error analysis by parts</a></h3>
<p><a href="chapter53.html">Chapter 53、 Error analysis by parts</a></p>
<p><a href="chapter54.html">Chapter 54、Attributing error to one part</a></p>
<p><a href="chapter55.html">Chapter 55、General case of error attribution</a></p>
<p><a href="chapter56.html">Chapter 56、Error analysis by parts and comparison to human-level performance</a></p>
<p><a href="chapter57.html">Chapter 57、Spotting a flawed ML pipeline</a></p>
<h3 id="conclusion"><a class="header" href="#conclusion">Conclusion</a></h3>
<p><a href="chapter58.html">Chapter 58、Building a superhero team - Get your teammates to read this</a></p>
<h2 id="英文原文"><a class="header" href="#英文原文">英文原文</a></h2>
<p>详见 draft 目录：</p>
<p>01-14章：<a href="draft/Ng_MLY01-01-14.pdf">Ng_MLY01-01-14.pdf</a></p>
<p>15-19章：<a href="draft/Ng_MLY02-15-19.pdf">Ng_MLY02-15-19.pdf</a></p>
<p>20-22章：<a href="draft/Ng_MLY03-20-22.pdf">Ng_MLY03-20-22.pdf</a></p>
<p>23-27章：<a href="draft/Ng_MLY04-23-27.pdf">Ng_MLY04-23-27.pdf</a></p>
<p>28-30章：<a href="draft/Ng_MLY05-28-30.pdf">Ng_MLY05-28-30.pdf</a></p>
<p>31-32章：<a href="draft/Ng_MLY06-31-32.pdf">Ng_MLY06-31-32.pdf</a></p>
<p>33-35章：<a href="draft/Ng_MLY07-33-35.pdf">Ng_MLY07-33-35.pdf</a></p>
<p>36-39章：<a href="draft/Ng_MLY08-36-39.pdf">Ng_MLY08-36-39.pdf</a></p>
<p>40-43章：<a href="draft/Ng_MLY09-40-43.pdf">Ng_MLY09-40-43.pdf</a></p>
<p>44-46章：<a href="draft/NG_MLY10-44-46.pdf">NG_MLY10-44-46.pdf</a></p>
<p>47-49章：<a href="draft/NG_MLY11-47-49.pdf">NG_MLY11-47-49.pdf</a></p>
<p>50-52章：<a href="draft/Ng_MLY12-50-52.pdf">Ng_MLY12-50-52.pdf</a></p>
<p>53-58章：<a href="draft/Ng_MLY13-53-58.pdf">Ng_MLY13-53-58.pdf</a></p>
</main>
<nav class="nav-wrapper" aria-label="Page navigation">
<!-- Mobile navigation buttons -->
<a rel="next" href="chapter1.html" class="mobile-nav-chapters next" title="Next chapter" aria-label="Next chapter" aria-keyshortcuts="Right">
<i class="fa fa-angle-right"></i>
</a>
<div style="clear: both"></div>
</nav>
</div>
</div>
<nav class="nav-wide-wrapper" aria-label="Page navigation">
<a rel="next" href="chapter1.html" class="nav-chapters next" title="Next chapter" aria-label="Next chapter" aria-keyshortcuts="Right">
<i class="fa fa-angle-right"></i>
</a>
</nav>
</div>
<!-- Livereload script (if served using the cli tool) -->
<script type="text/javascript">
var socket = new WebSocket("ws://localhost:3000/__livereload");
socket.onmessage = function (event) {
if (event.data === "reload") {
socket.close();
location.reload();
}
};
window.onbeforeunload = function() {
socket.close();
}
</script>
<script type="text/javascript">
window.playground_copyable = true;
</script>
<script src="elasticlunr.min.js" type="text/javascript" charset="utf-8"></script>
<script src="mark.min.js" type="text/javascript" charset="utf-8"></script>
<script src="searcher.js" type="text/javascript" charset="utf-8"></script>
<script src="clipboard.min.js" type="text/javascript" charset="utf-8"></script>
<script src="highlight.js" type="text/javascript" charset="utf-8"></script>
<script src="book.js" type="text/javascript" charset="utf-8"></script>
<!-- Custom JS scripts -->
</body>
</html>