-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathblog.html
130 lines (103 loc) · 4.86 KB
/
blog.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
<!DOCTYPE HTML>
<!--
Introspect by TEMPLATED
templated.co @templatedco
Released for free under the Creative Commons Attribution 3.0 license (templated.co/license)
-->
<html>
<head>
<title>Blog - MLLAB at SNU</title>
<meta charset="utf-8"/>
<meta name="viewport" content="width=device-width, initial-scale=1, user-scalable=no"/>
<link rel="stylesheet" href="assets/css/owl.carousel.css"/>
<link rel="stylesheet" href="assets/css/owl.theme.default.css"/>
<link rel="stylesheet" href="assets/css/main.css?v=0.075"/>
<script type="text/javascript">
var loc = window.location.href+'';
// Comment away this part when running python app.py
if (loc.indexOf('http://')==0){
window.location.href = loc.replace('http://','https://');
}
//
if (document.location.pathname == '/~hyunoh' || document.location.pathname.startsWith('/~hyunoh/')) {
document.location.pathname = '/hyunoh/';
}
</script>
</head>
<body>
<!-- Header -->
<header id="header">
<div class="inner header_inner">
<a href="." class="logo"><img class="logo-image" src="./images/snu_logo_white.png" /> <img class="logo-image" src="./images/mllab-logo-new.png" /></a>
<nav id="nav">
<a href="." class="">Home</a>
<a href="people.html" class="">People</a>
<a href="publications.html" class="">Publications</a>
<a href="courses.html" class="">Courses</a>
<a href="blog.html" class="current">Blog</a>
</nav>
</div>
</header>
<a href="#menu" class="navPanelToggle"><span class="fa fa-bars"></span></a>
<!-- Main -->
<section id="main">
<div class="inner">
<header class="major special">
<h1>Blog</h1>
</header>
<br/>
<section class="entry">
<span class="entry-theme-blog underbar">
<span class="entry-theme-blog-head">
<span class="entry-theme-blog-title">
<h2>Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixup</h2>
</span>
<span class="entry-theme-blog-date">
<p>August 27, 2020</p>
</span>
<span class="entry-theme-blog-author">
<p>Article by Jang-Hyun Kim and Wonho Choo</p>
</span>
</span>
<img class="image blog-image" src="./images/posts/kim-ICML2020/animation.gif"/>
<p>Deep neural networks become the bedrock of modern AI tasks such as object recognition, speech, natural language processing, and reinforcement learning. However, these models are known to memorize the training data and make overconfident predictions often resulting in degraded generalization performance on test examples [Zhang2016]. Data augmentation approaches aim to alleviate some of these issues by improving the model generalization performance [Bishop2006].</p>
<br/>
<p>Recently, a line of new data augmentation techniques called <em>mixup</em> has been proposed [Zhang2018]. These methods mainly focus on creating previously unseen virtual mixup examples by interpolating a pair of given data. In this article, we will cover these methods with our paper “Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal mixup”.</p>
<br/>
<span class="entry-theme-blog-foot">
<a href="kim-ICML2020.html" style="font-weight:bold">Read more</a>
</span>
</span>
</section>
<div class="pagination">
<!-- <a href="#">«</a> -->
<a href="blog.html">1</a>
<!--
<a href="#">2</a>
<a href="#">3</a>
<a href="#">4</a>
<a href="#">5</a>
<a href="#">6</a>
<a href="#">»</a> -->
</div>
</div>
</section>
<!-- Footer -->
<section id="footer">
<div class="inner">
<strong>MLLAB, Computer Science and Engineering, Seoul National University</strong>
</br>
<a class="icon fa-map-marker"><span class="label">Location</span></a> Samsung Electronics-Seoul National University Research Center (Building 944),<br> 1, Gwanak-ro, Gwanak-gu, Seoul 08826, Republic of Korea
<div class="copyright">
© SNU MLLAB Design: <a href="https://templated.co/">TEMPLATED</a>. Images <a
href="https://unsplash.com/">Unsplash</a>
</div>
</div>
</section>
<!-- Scripts -->
<!-- <script src="assets/js/jquery.min.js"></script>
<script src="assets/js/skel.min.js"></script>
<script src="assets/js/util.js"></script>
<script src="assets/js/main.js"></script> -->
</body>
</html>