-
Notifications
You must be signed in to change notification settings - Fork 72
/
index.html
359 lines (302 loc) · 17.6 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
<!doctype html>
<html lang="en" class="h-100">
<head>
<title>What are embeddings?</title>
<meta name="description" content="A deep-dive into machine learning embeddings. ">
<!-- Google / Search Engine Tags -->
<meta itemprop="name" content="machine learning, embeddings, deep learning, vector search">
<meta itemprop="description" content="A deep-dive into machine learning embeddings. ">
<meta itemprop="image"
content="https://raw.githubusercontent.com/veekaybee/what_are_embeddings/site/docs/assets/kandinsky.png">
<!-- Facebook Meta Tags -->
<meta property="og:url" content="http://vickiboykis.com/what_are_embeddings/index.html">
<meta property="og:type" content="website">
<meta property="og:title" content="What are embeddings?">
<meta property="og:description" content="A deep-dive into machine learning embeddings. ">
<meta property="og:image"
content="https://raw.githubusercontent.com/veekaybee/what_are_embeddings/site/docs/assets/kandinsky.png">
<!-- Twitter Meta Tags -->
<meta name="twitter:card" content="summary">
</meta>
<meta name="twitter:title" content="What are embeddings?">
</meta>
<meta name="twitter:description" content="A deep-dive into machine learning embeddings. ">
</meta>
<meta name="twitter:image"
content="https://raw.githubusercontent.com/veekaybee/what_are_embeddings/site/docs/assets/kandinsky.png">
</meta>
</meta>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link href="assets/dist/css/bootstrap.min.css" rel="stylesheet">
<style>
.bd-placeholder-img {
font-size: 1.125rem;
text-anchor: middle;
-webkit-user-select: none;
-moz-user-select: none;
user-select: none;
}
@media (min-width: 768px) {
.bd-placeholder-img-lg {
font-size: 3.5rem;
}
}
.b-example-divider {
height: 3rem;
background-color: rgba(0, 0, 0, .1);
border: solid rgba(0, 0, 0, .15);
border-width: 1px 0;
box-shadow: inset 0 .5em 1.5em rgba(0, 0, 0, .1), inset 0 .125em .5em rgba(0, 0, 0, .15);
}
.b-example-vr {
flex-shrink: 0;
width: 1.5rem;
height: 100vh;
}
.bi {
vertical-align: -.125em;
fill: rgb(221, 131, 132);
}
.nav-scroller {
position: relative;
z-index: 2;
height: 2.75rem;
overflow-y: hidden;
}
.nav-scroller .nav {
display: flex;
flex-wrap: nowrap;
padding-bottom: 1rem;
margin-top: -1px;
overflow-x: auto;
text-align: center;
white-space: nowrap;
-webkit-overflow-scrolling: touch;
}
.blue-div {
background-color: #a8cbfe;
;
}
.indigo-text {
color: #228c3a !important;
}
.left-align {
text-align: left;
}
.custom-padding {
padding-top: 20px;
}
img {
width: auto;
max-width: 100%;
height: auto;
}
td {
white-space: normal !important;
word-wrap: break-word;
}
</style>
<!-- Custom styles for this template -->
<link href="css/cover.css" rel="stylesheet">
</head>
<body class="text-center">
<div class="cover-container d-flex w-100 h-100 p-3 mx-auto flex-column">
<header class="mb-auto">
<div>
<nav class="nav nav-masthead justify-content-center text-primary float-md-end">
<a class="nav-link fw-bold py-1 px-0 active text-primary" aria-current="page"
href="https://www.vickiboykis.com/what_are_embeddings">Home</a>
<a class="nav-link fw-bold py-1 px-0 text-primary"
href="https://github.com/veekaybee/what_are_embeddings">GitHub</a>
</nav>
</div>
</header>
<main class="px-3">
<div class="container my-5">
<div class="p-5 text-center blue-div rounded-3">
<p><img src="assets/kandinsky.png" width="150" height="150"></p>
<h1 class="text-body-emphasis text-primary indigo-text">What are embeddings?</h1>
<p class="lead">
<div class="row">
<div class="col-md-12">
<a href="https://raw.githubusercontent.com/veekaybee/what_are_embeddings/main/embeddings.pdf"
class="btn btn-lg btn-success mb-2 mb-md-0">Get PDF</a>
<a href="https://github.com/veekaybee/what_are_embeddings" button type="button"
class="btn btn-lg btn-success">Get Code</a>
</div>
</div>
<script src="https://stackpath.bootstrapcdn.com/bootstrap/4.3.1/js/bootstrap.min.js"></script>
</p>
<ul class="nav nav-tabs gap-2 justify-content-center">
<li class="nav-item">
<a class="nav-link active" aria-current="page" href="index.html">Why</a>
</li>
<li class="nav-item">
<a class="nav-link" href="about.html">Who</a>
</li>
<li class="nav-item">
<a class="nav-link" href="next.html">Next</a>
</li>
</ul>
<div class="row custom-padding">
<div class="text-left gap-3">
<p class="left-align">If you go to the <a href="https://museupicassobcn.cat/en"> Museu
Picasso in Barcelona</a>, you'll
see many of the artist's early works. They're really interesting because they don't look
like what we think of as
Picasso's style. These paintings, completed during his early years, are
displays of his technical genius - as a classical painter. </p>
<p class="left-align">Some particularly amazing examples are "<a
href="https://en.wikipedia.org/wiki/Science_and_Charity">Science
and Charity</a>" and "<a
href="https://www.wikiart.org/en/pablo-picasso/first-communion-1896">First
Communion</a>". One of my favorites is "Portrait of the
Artist's Mother." These were all painted when he was fifteen. </p>
<p><img src="assets/portrait.png" width="300"></p>
<p class="left-align"> You can see
both the artist's
innate ability to make art, and his immense potential future. But to get to the point
where Picasso could reject traditional styles, he had
to master them first.
</p>
<p><img src="assets/yellow_hat.png" width="300"></p>
<p class="left-align"> This is also true for machine learning.
There's a
whole universe
of exciting developments at the forefront of large language models. But in the noise of
the bleeding edge, a lot of
important
foundational concepts get lost. If we don't
<a href="https://increment.com/software-architecture/architecture-for-generations/">
understand the fundamentals</a> of how we get from a single word to a BERT
representation, and more importantly, why we do so,
the models will remain black boxes to us. We won't be able to build on them and
master them in the ways
that we want.
</p>
<p class="left-align"> Peter Norvig urges us to <a
href="https://norvig.com/21-days.html">teach ourselves programming in ten
years</a>. In this spirit, after several years of working with embeddings,
foundational data structures in deep learning models, I realized it's not trivial to
have a good conceptual model of them. Moreover, when I did
want to learn more, there was no
good, general text I could refer to as a starting point. Everything was either too deep
and academic or too shallow and content from vendors in the space selling their
solution.
</a>
So I started a project to understand the
fundamental building blocks of machine learning and natural language processing,
particularly as they relate to recommendation systems today. The results of this
project are the PDF on this site, which is aimed at a generalist audience and not trying
to sell you anything except the idea that vectors are cool. I've also been working on <a
href="http://viberary.pizza">Viberary</a> to implement these ideas in practice.</p>
<p class=" left-align"> In addition to his art, Picasso also left us with the quote,
<blockquote class="blockquote"><svg xmlns="http://www.w3.org/2000/svg" width="16"
height="16" fill="currentColor" class="bi bi-chat-right-quote" viewBox="0 0 16 16">
<path
d="M12 12a1 1 0 0 0 1-1V8.558a1 1 0 0 0-1-1h-1.388c0-.351.021-.703.062-1.054.062-.372.166-.703.31-.992.145-.29.331-.517.559-.683.227-.186.516-.279.868-.279V3c-.579 0-1.085.124-1.52.372a3.322 3.322 0 0 0-1.085.992 4.92 4.92 0 0 0-.62 1.458A7.712 7.712 0 0 0 9 7.558V11a1 1 0 0 0 1 1h2Zm-6 0a1 1 0 0 0 1-1V8.558a1 1 0 0 0-1-1H4.612c0-.351.021-.703.062-1.054.062-.372.166-.703.31-.992.145-.29.331-.517.559-.683.227-.186.516-.279.868-.279V3c-.579 0-1.085.124-1.52.372a3.322 3.322 0 0 0-1.085.992 4.92 4.92 0 0 0-.62 1.458A7.712 7.712 0 0 0 3 7.558V11a1 1 0 0 0 1 1h2Z" />
</svg>
When art
critics
get together they talk about Form and Structure and Meaning. When artists get
together
they talk about where you can buy cheap turpentine.<svg
xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor"
class="bi bi-chat-left-quote" viewBox="0 0 16 16">
<path
d="M12 12a1 1 0 0 0 1-1V8.558a1 1 0 0 0-1-1h-1.388c0-.351.021-.703.062-1.054.062-.372.166-.703.31-.992.145-.29.331-.517.559-.683.227-.186.516-.279.868-.279V3c-.579 0-1.085.124-1.52.372a3.322 3.322 0 0 0-1.085.992 4.92 4.92 0 0 0-.62 1.458A7.712 7.712 0 0 0 9 7.558V11a1 1 0 0 0 1 1h2Zm-6 0a1 1 0 0 0 1-1V8.558a1 1 0 0 0-1-1H4.612c0-.351.021-.703.062-1.054.062-.372.166-.703.31-.992.145-.29.331-.517.559-.683.227-.186.516-.279.868-.279V3c-.579 0-1.085.124-1.52.372a3.322 3.322 0 0 0-1.085.992 4.92 4.92 0 0 0-.62 1.458A7.712 7.712 0 0 0 3 7.558V11a1 1 0 0 0 1 1h2Z" />
</svg></blockquote>
</p>
<p class=" left-align"> I wrote this text for my own learning process. But it's my
hope that this
document puts embeddings in a business and engineering context so that others including
engineers, PMs, students, and
anyone
looking to
learn more about
fundamentals finds it useful.</p>
<p class=" left-align">Machine learning, like all good engineering and like good art, is
ultimately, a way for us to express ourselves, a craft made up of fundamental building
blocks and patterns that empower us and
allow us to build something
beautiful on strong foundations of those that came before us. and I hope you find as
much joy in exploring embeddings and using them
as I did.</p>
<hr>
<a href="https://raw.githubusercontent.com/veekaybee/what_are_embeddings/main/embeddings.pdf"
class="btn btn-lg btn-success mb-2 mb-md-0">Get PDF</a>
<hr>
<p class="left-align">
<h2>Audience </h2>
<p class=" left-align">
Anyone who needs to understand embeddings, from machine learning engineers, to SWE,
product
or project manager, student, or simply curious. That said, there are different levels of
understanding in
the
document. Here are my recommendations on how to read:
<div class="table-responsive">
<table class="table table-bordered">
<thead class="thead-dark">
<tr>
<th scope="col">#</th>
<th scope="col">You want</th>
<th scope="col">Sections</th>
<th scope="col">Prereqs</th>
</tr>
</thead>
<tbody>
<tr>
<th scope="row">1</th>
<td>a high-level intro</td>
<td>1,2</td>
<td>how apps work </td>
</tr>
<tr>
<th scope="row">2</th>
<td>to understand engineering concerns</td>
<td>1,2,5</td>
<td>Python, engineering, Big O Notation</td>
</tr>
<tr>
<th scope="row">3</th>
<td> an ML internal deep dive</td>
<td>3,4</td>
<td>Linear Algebra, ML, Python, engineering</td>
</tr>
</tbody>
</table>
</div>
<hr>
<p class="left-align">
<h2>License </h2>
<p class="left-align">"What are embeddings" is licensed under a Creative Commons, By
Attribution, Non-Commercial, Share Alike 3.0 license. This means that you are welcome to
use, download and share the text, provided
that
attribution is given to the author (Vicki Boykis) and that it is released under the
same
license and <a href="https://creativecommons.org/licenses/by-nc-sa/3.0/us/">not used for
commercial purposes</a>.
</p>
<hr>
<p class="left-align">
<h2>Feedback </h2>
<p class="left-align">Feedback on the clarity of concepts or just typos is welcome. Feel
free to <a href="https://github.com/veekaybee/what_are_embeddings">submit a PR</a>
.
</p>
</div>
</div>
</div>
</div>
</main>
<hr>
</div></a>
<footer class="mt-auto text-black-50">
<p>A thing from <a href="https://vickiboykis.com/">Vicki Boykis</a>.</p>
</footer>
</div>
</body>
</html>