-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
73 lines (67 loc) · 2.08 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
<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html;charset=utf-8"/>
<title>Seein' In</title>
<link href='http://fonts.googleapis.com/css?family=EB+Garamond' rel='stylesheet' type='text/css'>
<style>
a {
color: #000;
text-decoration: none;
font-size: 120%;
display: inline-block;
margin: 0 auto;
width: 190px;
text-align: center;
}
a:hover {
color: #888;
}
body {
font-family: 'EB Garamond', serif;
}
div {
margin: 0 auto;
}
div.main {
max-width: 600px;
min-width: 600px;
}
h1 {
text-align: center;
font-size: 300%;
margin: 40px 0 0 0;
}
h3 {
text-align: center;
font-size: 150%;
margin: 0;
}
img {
margin: 0 auto;
display: block;
}
span {
text-align: center;
display: inline-block;
width: 15px;
}
</style>
</head>
<body>
<div class="main">
<h1>Seein' In</h1>
<h3>Tanner Schmidt</h3>
<hr/>
<div><a href="final/poster-tws10.pdf">Poster</a><span>|</span><a href="final/paper-tws10.pdf">Paper</a><span>|</span><a href="https://github.com/CSE512-15S/fp-tws10">Code</a></div>
<hr/>
<img src="images/fullWindow.png" width="600px"/>
<p><strong>Abstract: </strong>This paper presents Seein’ In, an interactive tool for visualizing trained convolutional neural networks. The core visualization
is a layer-by-layer scatter plot matrix, plotting pairs of dimensions of the intermediate feature maps produced by the network given
a large input dataset. This display is coordinated with another display showing at once all activations produced by a single image, or
the average activation produced by a subset of the images. A variety of navigation techniques support and encourage exploration of
the vast amount of data contained in these displays. This tool should prove useful for machine learning researchers wishing to make
more informed decisions about network architectures and learning parameters.</p>
</div>
</body>
</html>