forked from carpentries-incubator/deep-learning-intro
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdesign.html
523 lines (483 loc) · 31.1 KB
/
design.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
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
<!DOCTYPE html>
<!-- START: inst/pkgdown/templates/layout.html --><!-- Generated by pkgdown: do not edit by hand --><html lang="en" data-bs-theme="auto"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8"><meta charset="utf-8"><title>Introduction to deep learning: Lesson design</title><meta name="viewport" content="width=device-width, initial-scale=1"><script src="assets/themetoggle.js"></script><link rel="stylesheet" type="text/css" href="assets/styles.css"><script src="assets/scripts.js" type="text/javascript"></script><!-- mathjax --><script type="text/x-mathjax-config">
MathJax.Hub.Config({
config: ["MMLorHTML.js"],
jax: ["input/TeX","input/MathML","output/HTML-CSS","output/NativeMML", "output/PreviewHTML"],
extensions: ["tex2jax.js","mml2jax.js","MathMenu.js","MathZoom.js", "fast-preview.js", "AssistiveMML.js", "a11y/accessibility-menu.js"],
TeX: {
extensions: ["AMSmath.js","AMSsymbols.js","noErrors.js","noUndefined.js"]
},
tex2jax: {
inlineMath: [['\\(', '\\)']],
displayMath: [ ['$$','$$'], ['\\[', '\\]'] ],
processEscapes: true
}
});
</script><script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.5/MathJax.js" integrity="sha256-nvJJv9wWKEm88qvoQl9ekL2J+k/RWIsaSScxxlsrv8k=" crossorigin="anonymous"></script><!-- Responsive Favicon for The Carpentries --><link rel="apple-touch-icon" sizes="180x180" href="favicons/incubator/apple-touch-icon.png"><link rel="icon" type="image/png" sizes="32x32" href="favicons/incubator/favicon-32x32.png"><link rel="icon" type="image/png" sizes="16x16" href="favicons/incubator/favicon-16x16.png"><link rel="manifest" href="favicons/incubator/site.webmanifest"><link rel="mask-icon" href="favicons/incubator/safari-pinned-tab.svg" color="#5bbad5"><meta name="msapplication-TileColor" content="#da532c"><meta name="theme-color" media="(prefers-color-scheme: light)" content="white"><meta name="theme-color" media="(prefers-color-scheme: dark)" content="black"></head><body>
<header id="top" class="navbar navbar-expand-md top-nav incubator"><svg xmlns="http://www.w3.org/2000/svg" class="d-none"><symbol id="check2" viewbox="0 0 16 16"><path d="M13.854 3.646a.5.5 0 0 1 0 .708l-7 7a.5.5 0 0 1-.708 0l-3.5-3.5a.5.5 0 1 1 .708-.708L6.5 10.293l6.646-6.647a.5.5 0 0 1 .708 0z"></path></symbol><symbol id="circle-half" viewbox="0 0 16 16"><path d="M8 15A7 7 0 1 0 8 1v14zm0 1A8 8 0 1 1 8 0a8 8 0 0 1 0 16z"></path></symbol><symbol id="moon-stars-fill" viewbox="0 0 16 16"><path d="M6 .278a.768.768 0 0 1 .08.858 7.208 7.208 0 0 0-.878 3.46c0 4.021 3.278 7.277 7.318 7.277.527 0 1.04-.055 1.533-.16a.787.787 0 0 1 .81.316.733.733 0 0 1-.031.893A8.349 8.349 0 0 1 8.344 16C3.734 16 0 12.286 0 7.71 0 4.266 2.114 1.312 5.124.06A.752.752 0 0 1 6 .278z"></path><path d="M10.794 3.148a.217.217 0 0 1 .412 0l.387 1.162c.173.518.579.924 1.097 1.097l1.162.387a.217.217 0 0 1 0 .412l-1.162.387a1.734 1.734 0 0 0-1.097 1.097l-.387 1.162a.217.217 0 0 1-.412 0l-.387-1.162A1.734 1.734 0 0 0 9.31 6.593l-1.162-.387a.217.217 0 0 1 0-.412l1.162-.387a1.734 1.734 0 0 0 1.097-1.097l.387-1.162zM13.863.099a.145.145 0 0 1 .274 0l.258.774c.115.346.386.617.732.732l.774.258a.145.145 0 0 1 0 .274l-.774.258a1.156 1.156 0 0 0-.732.732l-.258.774a.145.145 0 0 1-.274 0l-.258-.774a1.156 1.156 0 0 0-.732-.732l-.774-.258a.145.145 0 0 1 0-.274l.774-.258c.346-.115.617-.386.732-.732L13.863.1z"></path></symbol><symbol id="sun-fill" viewbox="0 0 16 16"><path d="M8 12a4 4 0 1 0 0-8 4 4 0 0 0 0 8zM8 0a.5.5 0 0 1 .5.5v2a.5.5 0 0 1-1 0v-2A.5.5 0 0 1 8 0zm0 13a.5.5 0 0 1 .5.5v2a.5.5 0 0 1-1 0v-2A.5.5 0 0 1 8 13zm8-5a.5.5 0 0 1-.5.5h-2a.5.5 0 0 1 0-1h2a.5.5 0 0 1 .5.5zM3 8a.5.5 0 0 1-.5.5h-2a.5.5 0 0 1 0-1h2A.5.5 0 0 1 3 8zm10.657-5.657a.5.5 0 0 1 0 .707l-1.414 1.415a.5.5 0 1 1-.707-.708l1.414-1.414a.5.5 0 0 1 .707 0zm-9.193 9.193a.5.5 0 0 1 0 .707L3.05 13.657a.5.5 0 0 1-.707-.707l1.414-1.414a.5.5 0 0 1 .707 0zm9.193 2.121a.5.5 0 0 1-.707 0l-1.414-1.414a.5.5 0 0 1 .707-.707l1.414 1.414a.5.5 0 0 1 0 .707zM4.464 4.465a.5.5 0 0 1-.707 0L2.343 3.05a.5.5 0 1 1 .707-.707l1.414 1.414a.5.5 0 0 1 0 .708z"></path></symbol></svg><a class="visually-hidden-focusable skip-link" href="#main-content">Skip to main content</a>
<div class="container-fluid top-nav-container">
<div class="col-md-8">
<div class="large-logo">
<img id="incubator-logo" alt="Carpentries Incubator" src="assets/images/incubator-logo.svg"></div>
</div>
<div class="selector-container">
<div id="theme-selector">
<li class="nav-item dropdown" id="theme-button-list">
<button class="btn btn-link nav-link px-0 px-lg-2 dropdown-toggle d-flex align-items-center" id="bd-theme" type="button" aria-expanded="false" data-bs-toggle="dropdown" data-bs-display="static" aria-label="Toggle theme (auto)">
<svg class="bi my-1 theme-icon-active"><use href="#circle-half"></use></svg><i data-feather="chevron-down"></i>
</button>
<ul class="dropdown-menu dropdown-menu-end" aria-labelledby="bd-theme-text"><li>
<button type="button" class="btn dropdown-item d-flex align-items-center" data-bs-theme-value="light" aria-pressed="false">
<svg class="bi me-2 theme-icon"><use href="#sun-fill"></use></svg>
Light
<svg class="bi ms-auto d-none"><use href="#check2"></use></svg></button>
</li>
<li>
<button type="button" class="btn dropdown-item d-flex align-items-center" data-bs-theme-value="dark" aria-pressed="false">
<svg class="bi me-2 theme-icon"><use href="#moon-stars-fill"></use></svg>
Dark
<svg class="bi ms-auto d-none"><use href="#check2"></use></svg></button>
</li>
<li>
<button type="button" class="btn dropdown-item d-flex align-items-center active" data-bs-theme-value="auto" aria-pressed="true">
<svg class="bi me-2 theme-icon"><use href="#circle-half"></use></svg>
Auto
<svg class="bi ms-auto d-none"><use href="#check2"></use></svg></button>
</li>
</ul></li>
</div>
<div class="dropdown" id="instructor-dropdown">
<button class="btn btn-secondary dropdown-toggle bordered-button" type="button" id="dropdownMenu1" data-bs-toggle="dropdown" aria-expanded="false">
<i aria-hidden="true" class="icon" data-feather="eye"></i> Learner View <i data-feather="chevron-down"></i>
</button>
<ul class="dropdown-menu" aria-labelledby="dropdownMenu1"><li><button class="dropdown-item" type="button" onclick="window.location.href='instructor/design.html';">Instructor View</button></li>
</ul></div>
</div>
</div>
<hr></header><nav class="navbar navbar-expand-xl bottom-nav incubator" aria-label="Main Navigation"><div class="container-fluid nav-container">
<button class="navbar-toggler" type="button" data-bs-toggle="collapse" data-bs-target="#navbarSupportedContent" aria-controls="navbarSupportedContent" aria-expanded="false" aria-label="Toggle Navigation">
<span class="navbar-toggler-icon"></span>
<span class="menu-title">Menu</span>
</button>
<div class="nav-logo">
<img class="small-logo" alt="Carpentries Incubator" src="assets/images/incubator-logo-sm.svg"></div>
<div class="lesson-title-md">
Introduction to deep learning
</div>
<div class="search-icon-sm">
<!-- TODO: do not show until we have search
<i role="img" aria-label="Search the All In One page" data-feather="search"></i>
-->
</div>
<div class="desktop-nav">
<ul class="navbar-nav me-auto mb-2 mb-lg-0"><li class="nav-item">
<span class="lesson-title">
Introduction to deep learning
</span>
</li>
<li class="nav-item">
<a class="nav-link" href="key-points.html">Key Points</a>
</li>
<li class="nav-item">
<a class="nav-link" href="reference.html#glossary">Glossary</a>
</li>
<li class="nav-item">
<a class="nav-link" href="profiles.html">Learner Profiles</a>
</li>
<li class="nav-item dropdown">
<button class="nav-link dropdown-toggle" id="navbarDropdown" data-bs-toggle="dropdown" aria-expanded="false">
More <i data-feather="chevron-down"></i>
</button>
<ul class="dropdown-menu" aria-labelledby="navbarDropdown"><li><a class="dropdown-item" href="reference.html">Reference</a></li>
</ul></li>
</ul></div>
<!--
<form class="d-flex col-md-2 search-form">
<fieldset disabled>
<input class="form-control me-2 searchbox" type="search" placeholder="" aria-label="">
<button class="btn btn-outline-success tablet-search-button" type="submit">
<i class="search-icon" data-feather="search" role="img" aria-label="Search the All In One page"></i>
</button>
</fieldset>
</form>
-->
<a id="search-button" class="btn btn-primary" href="aio.html" role="button" aria-label="Search the All In One page">Search the All In One page</a>
</div><!--/div.container-fluid -->
</nav><div class="col-md-12 mobile-title">
Introduction to deep learning
</div>
<aside class="col-md-12 lesson-progress"><div style="width: NA%" class="percentage">
NA%
</div>
<div class="progress incubator">
<div class="progress-bar incubator" role="progressbar" style="width: NA%" aria-valuenow="NA" aria-label="Lesson Progress" aria-valuemin="0" aria-valuemax="100">
</div>
</div>
</aside><div class="container">
<div class="row">
<!-- START: inst/pkgdown/templates/navbar.html -->
<div id="sidebar-col" class="col-lg-4">
<div id="sidebar" class="sidebar">
<nav aria-labelledby="flush-headingEleven"><button role="button" aria-label="close menu" alt="close menu" aria-expanded="true" aria-controls="sidebar" class="collapse-toggle" data-collapse="Collapse " data-episodes="Episodes ">
<i class="search-icon" data-feather="x" role="img"></i>
</button>
<div class="sidebar-inner">
<div class="row mobile-row" id="theme-row-mobile">
<div class="col" id="theme-selector">
<li class="nav-item dropdown" id="theme-button-list">
<button class="btn btn-link nav-link px-0 px-lg-2 dropdown-toggle d-flex align-items-center" id="bd-theme" type="button" aria-expanded="false" data-bs-toggle="dropdown" data-bs-display="static" aria-label="Toggle theme (auto)">
<svg class="bi my-1 theme-icon-active"><use href="#circle-half"></use></svg><span class="d-lg-none ms-1" id="bd-theme-text">Toggle Theme</span>
</button>
<ul class="dropdown-menu dropdown-menu-right" aria-labelledby="bd-theme-text"><li>
<button type="button" class="btn dropdown-item d-flex align-items-center" data-bs-theme-value="light" aria-pressed="false">
<svg class="bi me-2 theme-icon"><use href="#sun-fill"></use></svg>
Light
<svg class="bi ms-auto d-none"><use href="#check2"></use></svg></button>
</li>
<li>
<button type="button" class="btn dropdown-item d-flex align-items-center" data-bs-theme-value="dark" aria-pressed="false">
<svg class="bi me-2 theme-icon"><use href="#moon-stars-fill"></use></svg>
Dark
<svg class="bi ms-auto d-none"><use href="#check2"></use></svg></button>
</li>
<li>
<button type="button" class="btn dropdown-item d-flex align-items-center active" data-bs-theme-value="auto" aria-pressed="true">
<svg class="bi me-2 theme-icon"><use href="#circle-half"></use></svg>
Auto
<svg class="bi ms-auto d-none"><use href="#check2"></use></svg></button>
</li>
</ul></li>
</div>
</div>
<div class="row mobile-row">
<div class="col">
<div class="sidenav-view-selector">
<div class="accordion accordion-flush" id="accordionFlush9">
<div class="accordion-item">
<h2 class="accordion-header" id="flush-headingNine">
<button class="accordion-button collapsed" id="instructor" type="button" data-bs-toggle="collapse" data-bs-target="#flush-collapseNine" aria-expanded="false" aria-controls="flush-collapseNine">
<i id="eye" aria-hidden="true" class="icon" data-feather="eye"></i> Learner View
</button>
</h2>
<div id="flush-collapseNine" class="accordion-collapse collapse" aria-labelledby="flush-headingNine" data-bs-parent="#accordionFlush2">
<div class="accordion-body">
<a href="instructor/design.html">Instructor View</a>
</div>
</div>
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
</div><!--div.sidenav-view-selector -->
</div><!--/div.col -->
<hr></div><!--/div.mobile-row -->
<div class="accordion accordion-flush" id="accordionFlush11">
<div class="accordion-item">
<button id="chapters" class="accordion-button show" type="button" data-bs-toggle="collapse" data-bs-target="#flush-collapseEleven" aria-expanded="false" aria-controls="flush-collapseEleven">
<h2 class="accordion-header chapters" id="flush-headingEleven">
EPISODES
</h2>
</button>
<div id="flush-collapseEleven" class="accordion-collapse show collapse" aria-labelledby="flush-headingEleven" data-bs-parent="#accordionFlush11">
<div class="accordion-body">
<div class="accordion accordion-flush" id="accordionFlush1">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading1">
<a href="index.html">Summary and Setup</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
<div class="accordion accordion-flush" id="accordionFlush2">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading2">
<a href="1-introduction.html">1. Introduction</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
<div class="accordion accordion-flush" id="accordionFlush3">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading3">
<a href="2-keras.html">2. Classification by a neural network using Keras</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
<div class="accordion accordion-flush" id="accordionFlush4">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading4">
<a href="3-monitor-the-model.html">3. Monitor the training process</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
<div class="accordion accordion-flush" id="accordionFlush5">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading5">
<a href="4-advanced-layer-types.html">4. Advanced layer types</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
<div class="accordion accordion-flush" id="accordionFlush6">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading6">
<a href="5-transfer-learning.html">5. Transfer learning</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
<div class="accordion accordion-flush" id="accordionFlush7">
<div class="accordion-item">
<div class="accordion-header" id="flush-heading7">
<a href="6-outlook.html">6. Outlook</a>
</div><!--/div.accordion-header-->
</div><!--/div.accordion-item-->
</div><!--/div.accordion-flush-->
</div>
</div>
</div>
<hr class="half-width"><div class="accordion accordion-flush lesson-resources" id="accordionFlush12">
<div class="accordion-item">
<h2 class="accordion-header" id="flush-headingTwelve">
<button class="accordion-button collapsed" id="lesson-resources" type="button" data-bs-toggle="collapse" data-bs-target="#flush-collapseTwelve" aria-expanded="false" aria-controls="flush-collapseTwelve">
RESOURCES
</button>
</h2>
<div id="flush-collapseTwelve" class="accordion-collapse collapse" aria-labelledby="flush-headingTwelve" data-bs-parent="#accordionFlush12">
<div class="accordion-body">
<ul><li>
<a href="key-points.html">Key Points</a>
</li>
<li>
<a href="reference.html#glossary">Glossary</a>
</li>
<li>
<a href="profiles.html">Learner Profiles</a>
</li>
<li><a href="reference.html">Reference</a></li>
</ul></div>
</div>
</div>
</div>
<hr class="half-width lesson-resources"><a href="aio.html">See all in one page</a>
<hr class="d-none d-sm-block d-md-none"><div class="d-grid gap-1">
</div>
</div><!-- /div.accordion -->
</div><!-- /div.sidebar-inner -->
</nav></div><!-- /div.sidebar -->
</div><!-- /div.sidebar-col -->
<!-- END: inst/pkgdown/templates/navbar.html-->
<!-- START: inst/pkgdown/templates/content-instructor.html -->
<div class="col-xl-8 col-lg-12 primary-content">
<nav class="lesson-content mx-md-4" aria-label="Previous and Next Chapter"><!-- content for small screens --><div class="d-block d-sm-block d-md-none">
<a class="chapter-link" href="index.html"><i aria-hidden="true" class="small-arrow" data-feather="arrow-left"></i>Previous</a>
<a class="chapter-link float-end" href="index.html">Next<i aria-hidden="true" class="small-arrow" data-feather="arrow-right"></i></a>
</div>
<!-- content for large screens -->
<div class="d-none d-sm-none d-md-block">
<a class="chapter-link" href="index.html" rel="prev">
<i aria-hidden="true" class="small-arrow" data-feather="arrow-left"></i>
Home
</a>
<a class="chapter-link float-end" href="index.html" rel="next">
Home
<i aria-hidden="true" class="small-arrow" data-feather="arrow-right"></i>
</a>
</div>
<hr></nav><main id="main-content" class="main-content"><div class="container lesson-content">
<h1>Lesson design</h1>
<p>Last updated on 2024-12-03 |
<a href="https://github.com/carpentries-incubator/deep-learning-intro/edit/main/instructors/design.md" class="external-link">Edit this page <i aria-hidden="true" data-feather="edit"></i></a></p>
<div class="text-end">
<button role="button" aria-pressed="false" tabindex="0" id="expand-code" class="pull-right" data-expand="Expand All Solutions " data-collapse="Collapse All Solutions "> Expand All Solutions <i aria-hidden="true" data-feather="plus"></i></button>
</div>
<p>This page documents the design process and motivation of this lesson
material.</p>
<p><strong>Lesson Title: An Introduction to Deep Learning</strong></p>
<section><h2 class="section-heading" id="target-audience">Target audience<a class="anchor" aria-label="anchor" href="#target-audience"></a></h2>
<hr class="half-width"><p>The main audience of this carpentry lesson is PhD students that have
little to no experience with deep learning. In addition, we expect them
to know basics of statistics and machine learning.</p>
<div class="section level3">
<h3 id="notes">Notes<a class="anchor" aria-label="anchor" href="#notes"></a></h3>
<ul><li>Probably have overhyped expectations of deep learning.</li>
<li>They don’t know if it’s the right tool for their situations.</li>
<li>They have no idea what it takes to actually do deep learning.</li>
<li>Want to quickly have some useful skills for their own data.</li>
</ul><div class="section level4">
<h4 id="required-pre-knowledge">Required Pre-Knowledge<a class="anchor" aria-label="anchor" href="#required-pre-knowledge"></a></h4>
<ul><li>
<strong>Python</strong> – Previous programming experience in Python
is required (Refer to Python Data Carpentry Lesson)</li>
<li>
<strong>Pandas</strong> – Knowledge of the Pandas Python
package</li>
<li>
<strong>Basic Machine Learning Knowledge</strong> – Data cleaning,
train & test split, overfitting & underfitting, metrics
(accuracy, recall, etc.),</li>
</ul></div>
</div>
</section><section><h2 class="section-heading" id="learning-objectives">Learning objectives<a class="anchor" aria-label="anchor" href="#learning-objectives"></a></h2>
<hr class="half-width"><blockquote>
<h2 id="overview">Overview</h2>
<p>After following this lesson, learners will be able to:</p>
<ul><li>Prepare input data for use for deep learning</li>
<li>Design and train a Deep Neural Network</li>
<li>Troubleshoot the learning process</li>
<li>Measure the performance of the network</li>
<li>Visualizing data and results</li>
<li>Re-use existing network architectures with and without pre-trained
weights</li>
</ul></blockquote>
<p>The following offers more details to each learning objective based on
Bloom’s Taxonomy. For hints on how to use this approach, see <a href="https://carpentries.github.io/instructor-training/18-preparation.html" class="external-link">lesson
18 of the instructor training</a></p>
<div class="section level3">
<h3 id="prepare-input-data-for-use-for-deep-learning">Prepare input data for use for deep learning<a class="anchor" aria-label="anchor" href="#prepare-input-data-for-use-for-deep-learning"></a></h3>
<p>This includes cleaning data, filling missing values, normalizing, and
transforming categorical columns into dummy encoding.</p>
<p>After this module, learners can …</p>
<ul><li>define a checklist for data analysis steps before applying Deep
Learning to the data</li>
<li>describe criteria by which to judge good or bad data, e.g. how a
column’s values should be distributed</li>
<li>execute a min-max normalization on floating point data</li>
<li>sketch how to insert missing timestamps or literal values
(i.e. factors or non-numeric entries)</li>
<li>implement a transformation of categorical values into a numerical
encoding (<code>int8</code>)</li>
<li>argue for or against strategies to normalize data</li>
<li>formulate techniques to prepare (clean) data for training a deep
learning network</li>
</ul></div>
<div class="section level3">
<h3 id="design-and-train-a-deep-neural-network">Design and train a Deep Neural Network<a class="anchor" aria-label="anchor" href="#design-and-train-a-deep-neural-network"></a></h3>
<p>This includes knowledge of when to different types of layers</p>
<p>After this module, learners can …</p>
<ul><li>list/repeat the three ingredients to a feed forward network: input,
hidden layers, output</li>
<li>classify/categorize parts of a feed forward network when presented a
network architecture (as from <code>keras.model.summary()</code>)</li>
<li>describe a fully connected (dense) layer</li>
<li>describe a convolutional layer</li>
<li>describe a max pooling layer</li>
<li>describe an activation function</li>
<li>describe a softmax layer</li>
<li>argue against abundant use of the sigmoid function
(exploding/vanishing gradients)</li>
<li>calculate the output data shape of an image when transformed by a
fixed convolutional layer</li>
<li>interpret errors with convolutional layers</li>
<li>execute a 3 layer network on the MNIST data (or similar)</li>
<li>differentiate a dense layer and a convolutional layer</li>
<li>experiment with values of dense layer and a convolutional layer</li>
<li>select a layer type depending on the input data</li>
<li>develop a 5 layer network that comprises both layer types</li>
</ul></div>
<div class="section level3">
<h3 id="monitoring-and-troubleshooting-the-learning-process">Monitoring and Troubleshooting the learning process<a class="anchor" aria-label="anchor" href="#monitoring-and-troubleshooting-the-learning-process"></a></h3>
<p>Often when designing neural networks training will not automatically
work very well. This requires setting the parameters of the training
algorithm correctly, modifying the design of the network or changing the
data pre-processing. After training, the performance of the network
should be checked to prevent overfitting.</p>
<p>After this module, learners can …</p>
<ul><li>define precision and recall/accuracy for a classification task</li>
<li>state that cross-validation is used in Deep Learning too</li>
<li>describe how to split a dataset into training/test/validation
set</li>
<li>describe how Drop-Out Layers work</li>
<li>execute a plot to draw the loss per epoch for training and test
set</li>
<li>compare values of precision and recall</li>
<li>differentiate a overfitting network from a well-behaved network</li>
<li>detect when a network is underfitting or overfitting</li>
<li>design countermeasures for overfitting (e.g. more dropout layers,
reduce model size)</li>
<li>design countermeasures for underfitting (e.g. larger model)</li>
<li>critique a provided network design</li>
</ul></div>
<div class="section level3">
<h3 id="visualizing-data-and-results">Visualizing Data and Results<a class="anchor" aria-label="anchor" href="#visualizing-data-and-results"></a></h3>
<p>Within each episode how to visualize data and results</p>
<p>After this module, learners can …</p>
<ul><li>identify important plots to create at the end of training (provide
selected samples and their prediction)</li>
<li>execute plotting of important variables during training (loss,
ROC)</li>
<li>use tensorboard and related callbacks during training</li>
<li>examine the results of a partners network</li>
<li>critique the results of a partners network</li>
</ul></div>
<div class="section level3">
<h3 id="re-use-existing-network-architectures-with-and-without-pre-trained-weights">Re-use existing network architectures with and without pre-trained
weights<a class="anchor" aria-label="anchor" href="#re-use-existing-network-architectures-with-and-without-pre-trained-weights"></a></h3>
<p>Re-use of architectures is common in deep learning. Especially when
using pre-trained weights (transfer-learning) it can also be very
powerful.</p>
<p>After this module, learners can …</p>
<ul><li>describe what transfer learning stands for</li>
<li>explain in what situations transfer learning is beneficial</li>
<li>solve common issues of transfer learning (such as different
resolutions of the original training and the training at hand)</li>
<li>test training under different data shape mitigation strategies</li>
<li>relate training time of a de-novo network and a pretrained one</li>
<li>relate prediction quality of a de-novo network and a pretrained
one</li>
</ul><!--
Place links that you need to refer to multiple times across pages here. Delete
any links that you are not going to use.
--></div>
</section></div> <!-- / div.lesson-content -->
</main><!-- / main#main-content.main-content --><nav class="bottom-pagination mx-md-4" aria-label="Previous and Next Chapter"><div class="d-block d-sm-block d-md-none">
<a class="chapter-link" href="index.html"><i aria-hidden="true" class="small-arrow" data-feather="arrow-left"></i>Previous</a>
<a class="chapter-link float-end" href="index.html">Next<i aria-hidden="true" class="small-arrow" data-feather="arrow-right"></i></a>
</div>
<!-- content for large screens -->
<div class="d-none d-sm-none d-md-block">
<a class="chapter-link" href="index.html" rel="prev">
<i aria-hidden="true" class="small-arrow" data-feather="arrow-left"></i>
Home
</a>
<a class="chapter-link float-end" href="index.html" rel="next">
Home
<i aria-hidden="true" class="small-arrow" data-feather="arrow-right"></i>
</a>
</div>
</nav></div> <!-- / div.primary-content.col-xs-12 -->
<!-- END: inst/pkgdown/templates/content-instructor.html-->
</div><!--/div.row-->
<footer class="row footer mx-md-3"><hr><div class="col-md-6">
<p>This lesson is subject to the <a href="CODE_OF_CONDUCT.html">Code of Conduct</a></p>
<p>
<a href="https://github.com/carpentries-incubator/deep-learning-intro/edit/main/instructors/design.md" class="external-link">Edit on GitHub</a>
| <a href="https://github.com/carpentries-incubator/deep-learning-intro/blob/main/CONTRIBUTING.md" class="external-link">Contributing</a>
| <a href="https://github.com/carpentries-incubator/deep-learning-intro/" class="external-link">Source</a></p>
<p><a href="https://github.com/carpentries-incubator/deep-learning-intro/blob/main/CITATION.cff" class="external-link">Cite</a> | <a href="mailto:[email protected]">Contact</a> | <a href="https://carpentries.org/about/" class="external-link">About</a></p>
</div>
<div class="col-md-6">
<p>Materials licensed under <a href="LICENSE.html">CC-BY 4.0</a> by the authors</p>
<p>Template licensed under <a href="https://creativecommons.org/licenses/by-sa/4.0/" class="external-link">CC-BY 4.0</a> by <a href="https://carpentries.org/" class="external-link">The Carpentries</a></p>
<p>Built with <a href="https://github.com/carpentries/sandpaper/tree/0.16.10" class="external-link">sandpaper (0.16.10)</a>, <a href="https://github.com/carpentries/pegboard/tree/0.7.7" class="external-link">pegboard (0.7.7)</a>, and <a href="https://github.com/carpentries/varnish/tree/1.0.5" class="external-link">varnish (1.0.5)</a></p>
</div>
</footer></div> <!-- / div.container -->
<div id="to-top">
<a href="#top">
<i class="search-icon" data-feather="arrow-up" role="img" aria-label="Back To Top"></i><br><!-- <span class="d-none d-sm-none d-md-none d-lg-none d-xl-block">Back</span> To Top --><span class="d-none d-sm-none d-md-none d-lg-none d-xl-block">Back</span> To Top
</a>
</div>
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "TrainingMaterial",
"@id": "https://carpentries-incubator.github.io/deep-learning-intro/design.html",
"inLanguage": "en",
"dct:conformsTo": "https://bioschemas.org/profiles/TrainingMaterial/1.0-RELEASE",
"description": "A Carpentries Lesson teaching foundational data and coding skills to researchers worldwide",
"keywords": "deep learning, keras, lesson, The Carpentries, neural networks",
"name": "Lesson design",
"creativeWorkStatus": "active",
"url": "https://carpentries-incubator.github.io/deep-learning-intro/design.html",
"identifier": "https://carpentries-incubator.github.io/deep-learning-intro/design.html",
"dateCreated": "2020-10-17",
"dateModified": "2024-12-03",
"datePublished": "2024-12-03"
}
</script><script>
feather.replace();
</script></body></html><!-- END: inst/pkgdown/templates/layout.html-->