forked from crbothe/discourse-wizard
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
269 lines (238 loc) · 14.9 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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<meta name="description" content="">
<meta name="author" content="Chandrakant Bothe" >
<title>Discourse-Wizard</title>
<link rel="shortcut icon" type="png" href="img/favicon.png">
<!-- Bootstrap core CSS -->
<link href="vendor/bootstrap/css/bootstrap.min.css" rel="stylesheet">
<!-- Custom fonts for this template -->
<link href="https://fonts.googleapis.com/css?family=Saira+Extra+Condensed:100,200,300,400,500,600,700,800,900" rel="stylesheet">
<link href="https://fonts.googleapis.com/css?family=Open+Sans:300,300i,400,400i,600,600i,700,700i,800,800i" rel="stylesheet">
<link href="vendor/font-awesome/css/font-awesome.min.css" rel="stylesheet">
<link href="vendor/devicons/css/devicons.min.css" rel="stylesheet">
<link href="vendor/simple-line-icons/css/simple-line-icons.css" rel="stylesheet">
<!-- Custom styles for this template -->
<link href="css/unspeech.css" rel="stylesheet">
<!-- highlight js -->
<link rel="stylesheet" href="css/default.min.css">
</head>
<body id="page-top">
<nav class="navbar navbar-expand-lg navbar-dark bg-primary fixed-top" id="sideNav">
<a class="navbar-brand js-scroll-trigger" href="#page-top">
<span class="d-block d-lg-none">Discourse-Wizard</span>
<span class="d-none d-lg-block">
<img class="img-fluid img-profile rounded-circle mx-auto mb-2" src="img/discourse_logo.png" alt="">
</span>
</a>
<button class="navbar-toggler" type="button" data-toggle="collapse" data-target="#navbarSupportedContent" aria-controls="navbarSupportedContent" aria-expanded="false" aria-label="Toggle navigation">
<span class="navbar-toggler-icon"></span>
</button>
<div class="collapse navbar-collapse" id="navbarSupportedContent">
<ul class="navbar-nav">
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#about" onclick="switchColors(this);">About</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#demo" onclick="switchColors(this);">Demonstration</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#news" onclick="switchColors(this);">News</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#models" onclick="switchColors(this);">Models</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#code" onclick="switchColors(this);">Code</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#refs" onclick="switchColors(this);">References</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#footprint" onclick="switchColors(this);">Footprint</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#ack" onclick="switchColors(this);">Acknowledgement</a>
</li>
<li class="nav-item">
<a style="color:black" class="nav-link js-scroll-trigger" href="#contact" onclick="switchColors(this);">Contact</a>
</li>
<!--li class="nav-item">
<a class="nav-link js-scroll-trigger" href="#cluster">Cluster</a>
</li>
<li class="nav-item">
<a class="nav-link js-scroll-trigger" href="#visualize">Visualize</a>
</li-->
</ul>
</div>
</nav>
<div class="container-fluid p-0">
<section class="unspeech-section p-3 p-lg-5 d-flex d-column" id="about" class="slide">
<div class="my-auto">
<h2 class="mb-0">
<span class="text-primary">Discourse-Wizard</span>
</h2>
<p class="mb-5">
Discourse-Wizard is to demonstrate the context in spoken language which comes from their sequential patterns. For example, a sentence is formed by a sequence of words, a conversation is formed by a sequence of utterances, and so on. Dialogue act represents a performative action of an utterance. There have been many approaches to model such a deep contextual-concept to analyse conversations. The neural approaches have been deployed and achieved state-of-the-art results. In this demonstration recurrent neural networks are used to model the context-based learning of dialogue acts.</p>
<h4>Same utterance can be used in different contexts</h4>
<object data="img/example.png" width="60%" align="left" border="0">
not supported in your browser
</object>
<p class="mb-5">
As you can see in the small piece of conversation example, Utt2 and Utt4 are very same (in this case they are same “Yeah”), however, the dialogue acts are very different. As they come from their context utterances Utt1 and Utt3 respectively. If it appears after a ‘Yes-No Question’ dialogue act it is more likely to have ‘Yes-Answer’ rather than ‘Backchannel’ or any other dialogue act. The presented example is from the Switchboard Dialogue Act (SwDA) Corpus, which is annotated with 42 such dialogue acts.
</p>
<h3>Features:</h3>
<ul>
<li>Discourse/conversational analysis</li>
<li>Dialogue act recognition at utterance level</li>
<li>Context-based dialogue act recognition, preceding utterances as a context</li>
</ul>
<!--h3>How it works:</h3-->
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="demo" class="slide">
<div class="my-auto">
<h2 class="mb-5">Demonstration</h2>
<h3> Discourse-Wizard:-
Dialogue Act Recognition Demo <h3>
<h6 class="mb-5">If you don't see the program below, it is due to the server down issue! <br>
if urgently needed please refer to <a class="js-scroll-trigger" href="#contact">contact</a> page.
</h6>
<iframe src="https://f4b2654d709d.eu.ngrok.io"
width="100%" height="850" frameborder="0" allowfullscreen sandbox="allow-forms allow-scripts">
<p> <a href="https">
Fallback link for browsers that don't support iframes. Click here. </a> </p>
</iframe>
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="news" class="slide">
<div class="my-auto">
<h2 class="mb-5">News:</h2>
<ul>
<li>March 2020 - Our <a href="https://arxiv.org/abs/1912.00819v3">recent paper </a>is accepted at <a href="https://lrec2020.lrec-conf.org/en/"> LREC 2020</a> under title "EDA: Enriching Emotional Dialogue Acts using an Ensemble of Neural Annotators".</li>
<li>January 2020 - Updated demo with <a href="https://arxiv.org/abs/1912.00819v2">new DA recognition system</a>.</li>
<li>December 2019 - <a href="https://arxiv.org/abs/1912.00819v2">A preprint of our new paper</a> on Automated Neural Ensemble Annotation DA Recongnizer is now available; with title <br/>"Enriching Existing Conversational Emotion Datasets with Dialogue Acts using Neural Annotators"</li>
<li>August 2018 - <a href="https://arxiv.org/pdf/1806.11420.pdf">A preprint of our paper</a> on Discourse-Wizard is now available; with title <br/>"Discourse-Wizard: Discovering Deep Discourse Structure in your Conversation with RNNs"</li>
<li>August 2018 - Code for the training and demonstraiton will be made availbale</li>
<li>26. June 2018 - Featured blog on Medium <br/> <a href="https://medium.com/@crbothe255/hierarchical-context-in-conversations-for-spoken-language-understanding-7149ce276556" target='_blank'>"Hierarchical Context in Conversations for Spoken Language Understanding"</a> </li>
<li>3. June 2018 - One of the related paper is accepted at <a href="http://interspeech2018.org/accepted-papers.html">Interspeech 2018!</a> <br/>"Conversational Analysis using Utterance-level Attention-based Bidirectional Recurrent Neural Networks" </li> <!-- Chandrakant Bothe, Cornelius Weber, Sven Magg and Stefan Wermter -->
<li>25. May 2018 - Initial release of the Discourse-Wizard live web-demo</li>
</ul>
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex d-column" id="models">
<div class="my-auto">
<h1 class="mb-0">
<!-- <span class="text-primary">Neural Models</span> -->
<h2 class="mb-5">Neural Models</h2>
</h1>
<br/>
<p class="mb-5"><br/>
The context-based approach, which can take the preceding utterance (or utterances) into account, is crucial for language understanding modules in any dialogue engines. We use the hierarchical recurrent neural networks (RNN) to model such a context for conversational analysis as shown in the followind slides.
</p>
<div class="subheading mb-5">Hierarchical Context-based Dialogue Act Recognition using RNNs</div>
<iframe src="https://docs.google.com/presentation/d/e/2PACX-1vT_dsWwC-KEbBOH6cdRNvCVMJMk2qYZBriECXppZkPqnQKwJT9Or-oabsyC6dfosNhySwpG7Fg5DOs3/embed?start=true&loop=true&delayms=3000" frameborder="0" width="700" height="423" allowfullscreen="true" mozallowfullscreen="true" webkitallowfullscreen="true"></iframe>
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="code">
<div class="my-auto">
<h2 class="mb-5">Code</h2>
The backend server API commands for this demo can be found at <a target="_blank" href="https://github.com/bothe/EDAs/">EDAs<a>, where you may find <a target="_blank" href="https://github.com/bothe/EDAs/blob/master/dialogue_act_annotator.py">dialogue_act_annotator.py<a> providing examples to help you annotate your own conversational data.
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="refs">
<div class="my-auto">
<h2 class="mb-5">References</h2>
<p><strong>Bothe, C.</strong>, Weber, C., Magg, S., and Wermter, S. (2020).<br/><strong>
<a data-htmlarea-external="1" target="_blank" href="https://arxiv.org/pdf/1912.00819v3.pdf">"EDA: Enriching Emotional Dialogue Acts using an Ensemble of Neural Annotators".</a></strong>
<br/>Proceedings of the Language Resources and Evaluation Conference (LREC-2020).</p>
<hr />
<p><strong>Bothe, C.</strong>, Weber, C., Magg, S., and Wermter, S. (2019).<br/><strong>
<a data-htmlarea-external="1" target="_blank" href="https://arxiv.org/pdf/1912.00819v2.pdf">"Enriching Existing Conversational Emotion Datasets with Dialogue Acts using Neural Annotators".</a></strong>
<br/>Final version above in the Proceedings of the Language Resources and Evaluation Conference (LREC-2020).</p>
<hr />
<p><strong>Bothe, C.</strong>, Magg, S., Weber, C., and Wermter, S. (2018).<br/><strong>
<a data-htmlarea-external="1" target="_blank" href="https://arxiv.org/pdf/1806.11420.pdf">Discourse-Wizard: Discovering Deep Discourse Structure in your Conversation with RNNs.</a></strong>
<br/>arXiv:1806.11420 [cs.CL]</p>
<hr />
<p><strong>Bothe, C.</strong>, Magg, S., Weber, C., and Wermter, S. (2018).<br/><strong>
<a data-htmlarea-external="1" target="_blank" href="https://arxiv.org/pdf/1805.06242.pdf">Conversational Analysis using Utterance-level Attention-based Bidirectional Recurrent Neural Networks.</a></strong>
<br/>Proceedings of INTERSPEECH 2018.</p>
<hr />
<p><strong>Bothe, C.</strong>, Weber, C., Magg, S., and Wermter, S. (2018).<br/><strong>
<a data-htmlarea-external="1" target="_blank" href="http://www.lrec-conf.org/proceedings/lrec2018/summaries/525.html">A Context-based Approach for Dialogue Act Recognition using Simple Recurrent Neural Networks. </a></strong>
<br/> Proceedings of the Language Resources and Evaluation Conference (LREC-2018).</p>
</p>
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="footprint">
<div class="my-auto">
<h2 class="mb-5">Footprint</h2>
<object data="img/map_27-07-19.png" width="100%">
not supported in your browser
</object>
Updated on 17.07.2019. Based on the user/visitor requests sent to the server. Generated by Google Map using <a target="_blank" href="https://github.com/vgm64/gmplot">gmplot<a>.
</p>
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="ack">
<div class="my-auto">
<h2 class="mb-5">Acknowledgement</h2>
"This project has received funding from the European Union’s Horizon 2020 framework programme for research and innovation under the Marie Sklodowska-Curie Grant Agreement No.642667 (<a target="_blank" href="https://secure-robots.eu/">SECURE<a>)."
<br/>
<a target="_blank" href="https://secure-robots.eu/">
<object data="img/secure_logo.png" href="https://secure-robots.eu/" height="100">
png not supported in your browser
</object><a>
<a target="_blank" href="https://ec.europa.eu/research/mariecurieactions/">
<object data="img/eu_logo.png" height="130" >
png not supported in your browser
</object><a>
<br/>
<br/>
The initial demonstration release is available at SECURE Website as <a target="_blank" href="https://secure-robots.eu/fellows/bothe/discourse-wizard-demo/">Discourse-Wizard Demo<a>, which is simply a replica of the <a class="js-scroll-trigger" href="#demo">Demonstration<a> page here.
</div>
</section>
<section class="unspeech-section p-3 p-lg-5 d-flex flex-column" id="contact">
<div class="my-auto">
<h2 class="mb-5">Contact</h2>
<br/>
If you want to leave feedback or contact the author please write back at bothe[at]informatik[dot]uni-hamburg[dot]de.
<br/>
</div>
</section>
</div>
<!-- Bootstrap core JavaScript -->
<script src="vendor/jquery/jquery.min.js"></script>
<script src="vendor/bootstrap/js/bootstrap.bundle.min.js"></script>
<script>
function changeColor(id)
{
document.getElementById(id).style.color = "#ff0000"; // forecolor
document.getElementById(id).style.backgroundColor = "#ff0000"; // backcolor
}
</script>
<!-- Plugin JavaScript -->
<script src="vendor/jquery-easing/jquery.easing.min.js"></script>
<script>
function myFunction() {
document.getElementById("changeColor").style.color = "red";
}
</script>
<!-- Custom scripts for this template -->
<script src="js/unspeech.min.js"></script>
<script type="text/javascript">
function switchColors(element)
{
links=document.getElementsByTagName("li") ;
for (var i = 0 ; i < links.length ; i ++)
links.item(i).style.color = 'black' ;
element.style.color='green' ;
}
</script>
<script src="js/highlight.min.js"></script>
<script>hljs.initHighlightingOnLoad();</script>
</body>
</html>