forked from xinlongj/xinlongj.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
comads.html
291 lines (215 loc) · 14.3 KB
/
comads.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
<!DOCTYPE html>
<html >
<head>
<!-- Site made with Mobirise Website Builder v5.3.0, https://mobirise.com -->
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="generator" content="Mobirise v5.3.0, mobirise.com">
<meta name="viewport" content="width=device-width, initial-scale=1, minimum-scale=1">
<link rel="shortcut icon" href="assets/images/site-1-128x128.png" type="image/x-icon">
<meta name="description" content="The details of the study on computational advertising">
<title>Computational Advertising</title>
<link rel="stylesheet" href="assets/bootstrap/css/bootstrap.min.css">
<link rel="stylesheet" href="assets/bootstrap/css/bootstrap-grid.min.css">
<link rel="stylesheet" href="assets/bootstrap/css/bootstrap-reboot.min.css">
<link rel="stylesheet" href="assets/tether/tether.min.css">
<link rel="stylesheet" href="assets/dropdown/css/style.css">
<link rel="stylesheet" href="assets/socicon/css/styles.css">
<link rel="stylesheet" href="assets/theme/css/style.css">
<link rel="preload" as="style" href="assets/mobirise/css/mbr-additional.css"><link rel="stylesheet" href="assets/mobirise/css/mbr-additional.css" type="text/css">
</head>
<body>
<!-- Analytics -->
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-10509316-3"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-10509316-3');
</script>
<!-- /Analytics -->
<section class="menu cid-rpQOWImaEh" once="menu" id="menu2-29">
<nav class="navbar navbar-expand beta-menu navbar-dropdown align-items-center navbar-fixed-top navbar-toggleable-sm">
<button class="navbar-toggler navbar-toggler-right" type="button" data-toggle="collapse" data-target="#navbarSupportedContent" aria-controls="navbarSupportedContent" aria-expanded="false" aria-label="Toggle navigation">
<div class="hamburger">
<span></span>
<span></span>
<span></span>
<span></span>
</div>
</button>
<div class="menu-logo">
<div class="navbar-brand">
<span class="navbar-caption-wrap"><a class="navbar-caption text-black display-4" href="https://mobirise.co">@ICT, CAS</a></span>
</div>
</div>
<div class="collapse navbar-collapse" id="navbarSupportedContent">
<ul class="navbar-nav nav-dropdown" data-app-modern-menu="true"><li class="nav-item">
<a class="nav-link link text-black display-4" href="index.html">
Home
</a>
</li>
<li class="nav-item">
<a class="nav-link link text-black display-4" href="calendar.html">
Calendar
</a>
</li>
<li class="nav-item">
<a class="nav-link link text-black display-4" href="resch.html">Research</a>
</li><li class="nav-item"><a class="nav-link link text-black display-4" href="pubs.html">Publications</a></li><li class="nav-item"><a class="nav-link link text-black display-4" href="teach.html">Teaching</a></li><li class="nav-item"><a class="nav-link link text-black display-4" href="stds.html">Students</a></li></ul>
<div class="navbar-buttons mbr-section-btn"><a class="btn btn-sm btn-primary display-4" href="https://scholar.google.com/citations?user=W8wrWfMAAAAJ&hl=en" target="_blank"><span class="socicon socicon-googlescholar mbr-iconfont mbr-iconfont-btn" style="font-size: 16px;"></span></a> <a class="btn btn-sm btn-primary display-4" href="mailto:jiangxinlong@ict.ac.cn"><span class="socicon socicon-mail mbr-iconfont mbr-iconfont-btn" style="font-size: 16px;"></span></a></div>
</div>
</nav>
</section>
<section once="footers" class="cid-rpRZOXFy6T" id="footer6-2a">
<div class="container">
<div class="media-container-row align-center mbr-white">
<div class="col-12">
<p class="mbr-text mb-0 mbr-fonts-style display-4">
© Last modified on Jan. 2021 by
</p>
</div>
</div>
</div>
</section>
<section class="mbr-section article content10 cid-rSkAurZj5i" id="content10-2b">
<div class="container">
<div class="inner-container" style="width: 100%;">
<hr class="line" style="width: 32%;">
<div class="section-text align-center mbr-white mbr-fonts-style display-5">Enlightenments, like accidents, happen only to prepared minds.
<div>--Herbert Simon</div></div>
<hr class="line" style="width: 32%;">
</div>
</div>
</section>
<section class="mbr-section content4 cid-rSkAvrdipV" id="content4-2c">
<div class="container">
<div class="media-container-row">
<div class="title col-12 col-md-8">
<h2 class="align-center pb-3 mbr-fonts-style display-2">Refinement for Online Advertisement</h2>
<h3 class="mbr-section-subtitle align-center mbr-light mbr-fonts-style display-5">machine learning methods for improving online advertisement recommendations, e.g. CTR/CVR predictions.</h3>
</div>
</div>
</div>
</section>
<section class="mbr-section article content9 cid-rSkAvGQZmC" id="content9-2d">
<div class="container">
<div class="inner-container" style="width: 100%;">
<hr class="line" style="width: 25%;">
<div class="section-text align-center mbr-fonts-style display-7">Light componentized models and approaches for improving online advertisement recommendations are devised in this project. We also strive to provide users with better personalized recommendation. Some of recent research processes are summarized as follows.</div>
<hr class="line" style="width: 25%;">
</div>
</div>
</section>
<section class="mbr-section content4 cid-sce0MBcRJA" id="content4-2x">
<div class="container">
<div class="media-container-row">
<div class="title col-12 col-md-8">
<h2 class="align-center pb-3 mbr-fonts-style display-5">Towards Explainable Conversational Recommendation (IJCAI 2020)</h2>
</div>
</div>
</div>
</section>
<section class="mbr-section content6 cid-sce0NbBVsq" id="content6-2y">
<div class="container">
<div class="media-container-row">
<div class="col-12 col-md-8">
<div class="media-container-row">
<div class="mbr-figure" style="width: 75%;">
<img src="assets/images/ijcai2020-331x172.png" alt="Mobirise" title="">
</div>
<div class="media-content">
<div class="mbr-section-text">
<p class="mbr-text mb-0 mbr-fonts-style display-7">In this paper, we introduce explainable conversational recommendation, which enables incremental improvement of both recommendation accuracy and explanation quality through multi-turn usermodel conversation. We design an incremental multi-task learning framework that enables tight collaboration between recommendation prediction, explanation generation, and user feedback integration. We also propose a multi-view feedback integration method to enable effective incremental model update. Empirical results demonstrate that our model not only consistently improves the recommendation accuracy but also generates explanations that fit user interests reflected in the feedbacks. <a href="https://www.ijcai.org/Proceedings/2020/0414.pdf" target="_blank">[paper]</a><br></p>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="mbr-section content4 cid-rSkAwgKsaH" id="content4-2e">
<div class="container">
<div class="media-container-row">
<div class="title col-12 col-md-8">
<h2 class="align-center pb-3 mbr-fonts-style display-5">Field-aware Calibration: A Simple and Empirically Strong
<div>Method for Reliable Probabilistic Predictions (WWW 2020)</div></h2>
</div>
</div>
</div>
</section>
<section class="mbr-section content6 cid-rSkAwsjW04" id="content6-2f">
<div class="container">
<div class="media-container-row">
<div class="col-12 col-md-8">
<div class="media-container-row">
<div class="mbr-figure" style="width: 85%;">
<img src="assets/images/neural-cali-490x243.png" alt="Mobirise" title="">
</div>
<div class="media-content">
<div class="mbr-section-text">
<p class="mbr-text mb-0 mbr-fonts-style display-7">In this paper, we introduce a new evaluation metric named field-level calibration error that measures the bias in predictions over the sensitive input field that the decision-maker concerns. We then propose Neural Calibration, a simple yet powerful post-hoc calibration method that learns to calibrate by making full use of the field-aware information over the validation set.<a href="https://dl.acm.org/doi/abs/10.1145/3366423.3380154" target="_blank">[paper]</a><br></p>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="mbr-section content4 cid-rSkAwEIq1Y" id="content4-2g">
<div class="container">
<div class="media-container-row">
<div class="title col-12 col-md-8">
<h2 class="align-center pb-3 mbr-fonts-style display-5">Warm Up Cold-start Advertisements: Improving CTR Predictions via Learning to Learn ID Embeddings (SIGIR 2019)</h2>
</div>
</div>
</div>
</section>
<section class="mbr-section content6 cid-rSkAwPWuFq" id="content6-2h">
<div class="container">
<div class="media-container-row">
<div class="col-12 col-md-8">
<div class="media-container-row">
<div class="mbr-figure" style="width: 70%;">
<img src="assets/images/cold-490x323.png" alt="Mobirise" title="">
</div>
<div class="media-content">
<div class="mbr-section-text">
<p class="mbr-text mb-0 mbr-fonts-style display-7">In this work, we aim at improving the performance of CTR predictions during both the cold-start phase and the warm-up phase. We propose an approach coined <strong>Meta-Embedding</strong> that learns how to learn better embeddings for new ad IDs to address the cold-start problem. Then the embedding generator trained by the method can also speed up the model fitting and take the place of trivial random initializer for new ID embeddings so as to warm up cold-start for the new ad. <a href="https://dl.acm.org/doi/10.1145/3331184.3331268" target="_blank">[paper]</a><a href="https://github.com/Feiyang/MetaEmbedding" target="_blank">[code]</a><br></p>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="mbr-section content4 cid-rSkAx8Q9Yy" id="content4-2i">
<div class="container">
<div class="media-container-row">
<div class="title col-12 col-md-8">
<h2 class="align-center pb-3 mbr-fonts-style display-5">Attention-driven Factor Model for Explainable Personalized Recommendation (SIGIR 2018, short)</h2>
</div>
</div>
</div>
</section>
<section class="mbr-section content6 cid-rSkAxl9Yy5" id="content6-2j">
<div class="container">
<div class="media-container-row">
<div class="col-12 col-md-8">
<div class="media-container-row">
<div class="mbr-figure" style="width: 70%;">
<img src="assets/images/777-490x382.png" alt="Mobirise" title="">
</div>
<div class="media-content">
<div class="mbr-section-text">
<p class="mbr-text mb-0 mbr-fonts-style display-7">In this work, we propose the Attention-driven Factor Model (<strong>AFM</strong>), which can not only integrate item features driven by users’ attention but also can give reasonable explanations for users’ preferences and keep a high prediction accuracy. Meanwhile, we use the Gated Attention Units to extract explicit users’ preference. Taking advantage of rating and item features, the algorithm considers the personalization of different users' attention, and shows good efficiency and accuracy in experiments. <a href="https://dl.acm.org/doi/10.1145/3209978.3210083" target="_blank">[paper]</a></p>
</div>
</div>
</div>
</div>
</div>
</div>
</section><section style="background-color: #fff; font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', 'Roboto', 'Helvetica Neue', Arial, sans-serif; color:#aaa; font-size:12px; padding: 0; align-items: center; display: flex;"><a href="https://mobirise.site/s" style="flex: 1 1; height: 3rem; padding-left: 1rem;"></a><p style="flex: 0 0 auto; margin:0; padding-right:1rem;"><a href="https://mobirise.site/d" style="color:#aaa;">This web page</a> was started with Mobirise</p></section><script src="assets/web/assets/jquery/jquery.min.js"></script> <script src="assets/popper/popper.min.js"></script> <script src="assets/bootstrap/js/bootstrap.min.js"></script> <script src="assets/tether/tether.min.js"></script> <script src="assets/smoothscroll/smooth-scroll.js"></script> <script src="assets/dropdown/js/nav-dropdown.js"></script> <script src="assets/dropdown/js/navbar-dropdown.js"></script> <script src="assets/touchswipe/jquery.touch-swipe.min.js"></script> <script src="assets/theme/js/script.js"></script>
</body>
</html>