-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathreservoirs.html
341 lines (293 loc) · 9.63 KB
/
reservoirs.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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Reservoir Computing</title>
<style>
body {
font-family: Arial, sans-serif;
background-color: #f0f0f0;
margin: 0;
padding: 0;
}
.header {
text-align: center;
padding: 20px;
}
.header img {
width: 200px;
height: 200px;
border-radius: 50%;
}
.header h1 {
margin-top: 10px;
margin-bottom: 0;
}
.header p {
margin-top: 0;
}
.nav {
text-align: center;
background-color: #333;
padding: 10px;
}
.nav a {
color: #fff;
text-decoration: none;
padding: 0 10px;
}
.nav a:hover {
text-decoration: underline;
}
.content {
padding: 20px;
}
.content {
padding: 20px;
background-color: #ffffff; /* this is the color white */
}
body {
font-family: Arial, sans-serif;
background-color: #f0f0f0;
margin: 0;
padding: 0;
}
.section {
display: flex;
align-items: center;
justify-content: space-around;
margin: 30px auto;
padding: 40px;
background-color: #fff;
border-radius: 5px;
box-shadow: 0 0 15px rgba(0,0,0,0.1);
width: 90%;
max-width: 1200px;
}
.section .img1 {
width: 100px;
height: 150px;
}
.section .img2 {
width: 450px;
height: 200px;
}
.section .img3 {
width: 500px;
height: 300px;
}
.section .img4 {
width: 100px;
height: 100px;
}
.section div {
margin-left: 40px;
width: 70%;
}
.section:nth-child(odd) div {
margin-right: 40px;
margin-left: 0;
}
.section:nth-child(odd) {
flex-direction: row-reverse;
}
.intro {
display: flex;
/*align-items: center;*/
/*justify-content: center;*/
margin: 30px auto;
padding: 30px;
background-color: #333;
color: #fff;
border-radius: 5px;
width: 90%;
max-width: 1200px;
/*text-align: center;*/
}
.intro a {
color: #FFD700; /* Change the color of links */
text-decoration: none; /* Remove the underline */
padding: 0px; /* Add some space around the links */
}
.intro a:hover {
color: #fff; /* Change the color when hovered */
text-decoration: underline; /* Underline when hovered */
}
.intro a:first-child {
padding-left: 0;
padding-right: 0;
}
/* CSS for Screen Width up to 600px (Mobile) */
@media only screen and (max-width: 600px) {
.header img {
width: 100px;
height: 100px;
}
.section {
flex-direction: column;
align-items: center;
padding: 20px;
width: 100%;
}
.section .img1, .section .img2, .section .img3, .section .img4 {
width: 100px;
height: auto;
}
.section div {
margin-left: 0;
margin-right: 0;
}
.intro {
flex-direction: column;
align-items: center;
padding: 20px;
width: 100%;
}
}
/* CSS for Screen Width up to 600px (Mobile) */
@media only screen and (max-width: 600px) {
.header img {
width: 100px;
height: 100px;
}
.section {
flex-direction: column;
align-items: center;
padding: 20px;
width: 100%;
}
.section .img1, .section .img2, .section .img3 {
width: 100px;
height: auto;
}
.section div {
margin-left: 0;
margin-right: 0;
}
.intro {
flex-direction: column;
align-items: center;
padding: 20px;
width: 100%;
}
}
.section .img1, .section .img2, .section .img3, .section .img4 {
width: 40%;
height: auto;
}
h1, h2, h4, h5, h6 {
text-align: center;
}
.image-background .esnimg {
background-color: #000;
display: flex;
justify-content: center;
align-items: center;
padding: 0px;
margin: auto;
}
.image-background .esnimg {
max-width: 50%;
height: auto;
}
/* Add styles for the .content paragraph */
.content p {
max-width: 800px;
margin: auto;
padding: 15px;
background-color: rgba(0,0,0,5%);
border-radius: 3px;
}
.content article {
max-width: 900px;
margin: auto;
}
/* Add this code */
.content {
max-width: 1000px;
margin: 0 auto;
}
.content p {
background-color: #f2f2f2; /* Pale gray */
padding: 15px;
border-radius: 5px;
}
.content .paragraph {
background-color: #f2f2f2; /* Pale gray */
padding: 15px;
border-radius: 5px;
margin-bottom: 15px; /* gives some space between paragraphs */
max-width: 800px;
margin-left: auto;
margin-right: auto;
}
.content .paragraph p {
margin: 0;
}
</style>
<link rel="stylesheet" href="style.css">
</head>
<body>
<div class="header">
<img src="portrait.jpg" alt="Your headshot">
<h1>Anthony Polloreno, Ph.D.</h1>
<h2>Research Engineer</h2>
</div>
<div class="nav">
<a href="index.html">Home</a> |
<a href="research.html">Research</a> |
<a href="engineering.html"> Engineering </a> |
<a href="https://twitter.com/ampolloreno">Twitter</a> |
<a href="https://www.linkedin.com/in/ampolloreno/">LinkedIn</a> |
<a href="https://github.com/ampolloreno">GitHub</a> |
<a href="papers.html">Papers</a> |
<a href="resume.pdf">Resume</a>|
<a href="reservoirs.html">Reservoir Computing I</a> |
<a href="reservoirs2.html">Reservoir Computing II</a>|
<a href="reservoirs3.html">Reservoir Computing III</a>
</div>
<div class="content" id="blogPost">
<article>
<h1>The Impact of Noise on Recurrent Neural Networks I</h1>
<div class="paragraph">
<p> In this post, we are going to study reservoir computing, a subset of recurrent neural networks that are used for
complex temporal processing tasks. We will consider the setting where our computational reservoir is subject to noise
and errors. A classic example of reservoir computing is in prediction of dynamics of physical systems,
however by using typical encodings as in transformer-based methods they can also be used to predict language
patterns for human-like conversation based on previous inputs.
</p>
<p>
For this post, there are at least two motivations behind considering reservoir computers as opposed to something like a transformer.
The first is academic. Many simply have not heard of reservoir computers. Second, reservoir computing is simple, which makes
it an excellent pedagogical tool. The training, as we will see, involves a simple linear layer, and the recurrent structure
is given by a random matrix. This simplicity lends itself to intepretability.
</p>
<p>
The specific kind of reservoir computer we are going to consider are echo state networks (ESNs). They are a very simple
network with a few tunable parameters: the network sparsity, the bleedthrough between time steps, an encoding map, a decoding map,
an internal transition map, and the size of the reservoir. An interesting property of ESNs is that by randomly
initializing the weights in the encoding map and internal transition map, it is possible to learn to predict time series
by simply training the decoding map, which can be taken to be linear, and is hence computable via a standard least squares estimator.
For the purposes of this tutorial, we will fix the sparsity and bleedthrough.
</p>
<p>
The goal of this tutorial is to begin to understand how noise affects the performance of reservoir computers as a function of
their size. Our analysis is more generally applicable to analog systems where the state of the system is a continuous
value, such as the neural activations of our reservoir. What we aim to demonstrate is that in this scenario,
noise tends to substantially impair the kinds of computations a system is able to perform. We start with a simple
notebook that introduces the model (echo state networks), and the computational task (NARMA10).
</p>
</div>
<iframe src="Webpost1.html" style="height:7650px; width:100%; border:none;" ></iframe>
<div class="paragraph">
<p>Check out the next notebook <a href="http://ampolloreno.com/reservoirs2.html">here</a>!</p>
<section style="text-align: center;">
<h3>Acknowledgements</h3>
<p>A special thanks to <a href="https://ohaithe.re/">Alex Meiburg</a>, <a href="https://andremelo.org/">André Melo</a> and <a href="https://chromotopy.org/">Eric Peterson</a> for feedback on this post!</p>
</section>
</article>
</div>
</body>
</html>