Talks/AI_PWD.html at gh-pages · mpnkhan/Talks · GitHub

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
<!doctype html>
<html lang="en">

	<head>
		<meta charset="utf-8">

		<title>How AI can empower disabled people</title>

		<meta name="description" content="How AI can empower disabled people">
		<meta name="author" content="Prem Nawaz Khan">

		<meta name="apple-mobile-web-app-capable" content="yes">
		<meta name="apple-mobile-web-app-status-bar-style" content="black-translucent">

		<meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no">

		<link rel="stylesheet" href="reveal.js/dist/reset.css">
		<link rel="stylesheet" href="reveal.js/dist/reveal.css">
		<link rel="stylesheet" href="reveal.js/dist/theme/sky.css" id="theme">

		<!-- Theme used for syntax highlighting of code -->
		<link rel="stylesheet" href="reveal.js/lib/css/zenburn.css">
		<style type="text/css">
		ul{ font-size:80% !important }
		img{border: 0 !important}
		</style>
		<!-- Printing and PDF exports -->
		<script>
			var link = document.createElement( 'link' );
			link.rel = 'stylesheet';
			link.type = 'text/css';
			link.href = window.location.search.match( /print-pdf/gi ) ? 'reveal.js/css/print/pdf.css' : 'reveal.js/css/print/paper.css';
			document.getElementsByTagName( 'head' )[0].appendChild( link );
		</script>

		<!--[if lt IE 9]>
		<script src="lib/js/html5shiv.js"></script>
		<![endif]-->
	</head>

	<body>

		<div class="reveal">

			<!-- Any section element inside of this container is displayed as a slide -->
			<div class="slides">

			<section>
					<h2>How AI can empower disabled people</h2>
					<img src="imgs/GAAD_India_logo.jpg" alt="Global Accessibility Awareness Day, India" style="height:100px; width:100px;border:0px;">
					<h3>Global Accessibility Awarness Day</h3>
					<h4>Thursday, 17<sup>th</sup> of May 2018</h4>
					<p>

						<small>By Prem Nawaz Khan / <a href="http://twitter.com/mpnkhan">@mpnkhan</a></small>
					</p>
				</section>
				<section>
					<h2>About Me</h1>
					<img src="PayPal/imgs/common/Evangelist.jpg" alt="code, preach at PayPal" style="max-height:500px">
				</section>
				<section>
					<p> Statistics </p>
					<ul>
						<li>
							Assistive products: Hearing aids, wheelchairs, communication aids, spectacles, prostheses, pill organizers and memory aids .
						</li>
						<li>
							Globally, more than 1 billion people need 1 or more assistive products.
						</li>
						<li>
						More than 2 billion people will need at least 1 assistive product by 2050, <br>
						with many older people needing 2 or more.
						</li>
						<li>
							<b>Today, only 1 in 10 people in need have access to assistive products.</b>
						</li>
					</ul>
					<p>
						<small>References:</small>
						<small><a href="http://www.who.int/en/news-room/fact-sheets/detail/assistive-technology">World Health Organization Report</a> ,</small>
						<small><a href="https://en.wikipedia.org/wiki/World_report_on_disability" target="main">WRD Report</a></small>
					</p>
				</section>
				<section>
					<h3> Artifical Intelligence can empower people with disabilities</h3>
					<ul>
						<li> Voice Recognition </li>
						<li> Image recognition </li>
						<li> Facial recognition </li>
						<li> Lip-reading recognition </li>
						<li> NLP for Learners with Autism or Language Disorders</li>
						<li> Text to Sign Language Interpretation </li>
					</ul>
				</section>
				<section>
					<h2>Voice Recognition</h2>
					<p> Smart phones &  Smart speakers - best examples of reviving AI </p>
					<p>Apple's Siri, Amazon echo, Google Assistant, Cortana for Windows</p>
					<p> Pros & Cons </p>
					<aside class="notes">The field of AI research was born at a workshop at Dartmouth College in 1956</aside>
				</section>
				<section>
					<h3>Google Assistant, Duplex demo at Google IO May 8, 2018</h3>
					  <video controls tabIndex="0">
					    <source data-src="PayPal/imgs/AI/GoogleAssistant_Haircut_Appointment.mp4" type="video/mp4" />
					  </video>
				</section>
				<!--section>
					<h3>Bad:: Voice Only Recognition</h3>
					<img src="PayPal/imgs/AI/Text_to_SmartDevice_for_voiceless.png" alt="Need of Text to voice only devices" style="max-height:500px">
					<aside class="notes">Apple Siri and Google Assistant on smart phones has text to siri etc</aside>
				</section-->
				<section>
					<h3>Image recognition</h3>
					<ul>
						<li>Since 2012, Search engine had  deep learning solutions for content filtering and Image recognition </li>
						<li>But adding it for People with Disabilities helps the blind in recognizing images </li>
					</ul>
					<a href="https://github.com/yahoo/open_nsfw">
						<img src="PayPal/imgs/AI/Yahoo_ NSFW.png" alt="Not suitable for work (NSFW) image detection by Yahoo">
					</a>
				</section>
				<section>
					<h3>Alternate text, biggest barrier for PWD</h3>
					<pre><code class="html" data-trim contenteditable style="font-size: 18px;">
<img src="niceFlower.jpg"  alt="Lily in a Pond">
					</code></pre>
				</section>
				<section>
					<h3>For websites like PayPal, </h3>
					<p> It is about educating Content editors to input alternate text in their Content Management System</p>(and/or)
					<p> Educating Web Developers to put alt text for images</p>
				</section>
				<section>
					<ul>
						<li> Every day, people share more than 2 billion photos across Facebook, Instagram, Messenger, and WhatsApp </li>
						<li> Facebook announced their automatic alt text in April 2016</li>
					</ul>
						<img src="PayPal/imgs/AI/FB_AutoAlt.jpg" alt="automatic detection of Facebook">
					</ul>
				</section>
				<section>
					<p>April 2017, Automatic Alt text extension for Chrome </p>
						<img src="PayPal/imgs/AI/Chromext_autoalt.png" alt="automatic alt text chrome extension">
					<p> Based on Tensor flow and im2txt Model</p>
				</section>
				<section>
					<p>April 2018, Around 25 API's made available by Microsoft for public use</p>

						<img src="PayPal/imgs/AI/MS_CognitiveAPIs.png" alt="Microsoft API's">
					</ul>
				</section>
				<section>
					<h2>Image Alt text Demo</h2>
					<!-- <a href="http://localhost:3000/alttext.html" target="main">Demo</a> -->
					<a href="https://mpnkhan.github.io/newtech/alttext.html" target="main">Demo</a>

				</section>
				<section>
					<h3>Facial recognition</h3>
					<p>Using Facial recognition, we can</p>
					<ul>
						<li>Search, identify, and match faces in your private repository</li>
						<li>Find Similar faces</li>
						<li>Organize many unidentified faces together into groups, based on their visual similarity</li>
					</ul>
				</section>
				<section>
					<h3>Facial recognition</h3>
					<p>Recognizing Humans versus Bots</p>
					<img src="https://c.s-microsoft.com/en-in/CMSImages/Windows_Hello_1920_Carousel_en-US.jpg?version=33606bc2-f95f-b494-27df-28c7249f876f" alt="facial replaces captcha" style="max-height:300px">
					<p>CAPTCHA sucks?<br>
					Examples like <a href="https://support.apple.com/en-in/HT208108">iPhone X Facial Unlock</a>, <a href="https://www.microsoft.com/en-in/windows/windows-hello">Windows Hello</a>
					</p>
				</section>
				<section>
					<h2>Facial recognition Demo</h2>
					<!-- <a href="http://localhost:3000/cognitiveapi/" target="main">Demo</a> -->
					<a href="https://mpnkhan.github.io/newtech/cognitiveapi/" target="main">Demo</a>

				</section>
				<section>
					<h3>Lip-reading recognition</h3>
					<p>Automated captions using Lip Reading</p>
					<iframe width="854" height="480" src="https://www.youtube.com/embed/jWvvz0Xh0uU" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
<!--
					  <video controls tabIndex="0">
					    <source data-src="PayPal/imgs/AI/Lip_segmentation.mp4" type="video/mp4" />
					  </video>
-->
				</section>
				<section >
					<p>Oxford University researchers partnered with Google on a new AI tool that reads lips, and the results were significant. <br>
					<img src="PayPal/imgs/AI/this-is-how-deepmind-learned-to-read-lips.jpeg" alt="Google Deep Mind lip reading">
					 Trained with a dataset of more than 100,000 natural sentences.
					</p>
					<small>
						Source: <a href="https://www.techrepublic.com/article/google-deepmind-ai-destroys-human-expert-in-lip-reading-competition/">Tech Republic</a>
					, <a href="https://arxiv.org/pdf/1611.05358v1.pdf">Research Paper on Lip reading sentences in the Wild</a>
					</small>
				</section>
				<section>
					<h3>NLP (Natural language processing) for Learners with Autism or Language Disorders</h3>
					<p><a href="https://play.google.com/store/apps/details?id=com.equadex.helpicto&hl=en">Helpicto</a> is an android app which uses speech to text and Microsoft Cognitive API to convert speech a set of images which students with Language disorders related to autism, dysphasia, or Alzheimer’s disease</p>
				</section>
				<section>
					<p>Cognitive API uses AI to split the sentences and sends back the intents which is converted to a list of images<br>
						The speech command is “Do you want to eat an apple?” Helpicto will then generate three images: the child himself, the action of eating, and a picture of an apple.
					</p>
					<small>
						More info: <a href="https://microsoft.github.io/techcasestudies/cognitive%20services/2017/08/04/equadexcognitives.html/">How Equadex used Cognitive Services to help people with language disorders</a>
					</small>
				</section>
				<section>
					<h3>Text to Sign Language Interpretation</h3>
					<p>Real Time American Sign Language Video Captioning using Deep Neural Networks
					<img src="PayPal/imgs/AI/sign_language.png" alt="Project to convert text to Sign Language" style="max-height:300px">
					</p>

					<small>
						More info: <a href="http://on-demand.gputechconf.com/gtc/2017/presentation/s7346-syed-ahmed-real-time-american-sign-language-video-caption.pdf">Slides</a> and
						<a href="https://blogs.nvidia.com/blog/2017/05/11/ai-translates-sign-language/">NVIDIA Blog</a>
					</small>
				</section>
				<section>
					<h3>How you can help?</h3>
					<p> As a Front End Engineer </p>
					<p> Make use of WebRTC in detecting
						<a href="https://webgazer.cs.brown.edu/">Eye</a>,
						<a href="https://github.com/mtschirs/js-objectdetect/tree/gh-pages/examples">Hands</a>,
						<a href="https://github.com/auduno/clmtrackr">Face</a>
					</p>
					<p>Use Web Speech API</p>
					<pre><code class="html" data-trim contenteditable style="font-size: 18px;">
function synthVoice(text) {
  const synth = window.speechSynthesis;
  const utterance = new SpeechSynthesisUtterance();
  utterance.text = text;
  synth.speak(utterance);
}
					</code></pre>


					<p>SpeechRecognition API</p>
					<pre><code class="html" data-trim contenteditable style="font-size: 18px;">
const SpeechRecognition = window.SpeechRecognition || window.webkitSpeechRecognition;
const recognition = new SpeechRecognition();
					</code></pre>
					</p>
				</section>
				<section>
					<h4>How you can help?</h4>
					<p> As a Back End Engineer </br>
Open source Machine Learning Frameworks <br>
<a href="https://github.com/tensorflow">Google's Tensor flow </a><br>
<a href=" https://github.com/Microsoft/CNTK">Microsoft Cognitive Toolkit (CNTK) </a><br>
<a href="https://github.com/Theano/Theano">Theano</a><br>
<a href="https://www.mathworks.com/campaigns/products/trials/targeted/dpl.html">MATLAB</a>,
<a href="http://www.caffe2.ai">Caffe and Caffe2</a>,
<a href="https://keras.io/">Keras</a>,
<a href="https://github.com/pytorch/pytorch">Pytorch</a>,
<a href="https://github.com/apache/incubator-mxnet">MXNet</a>,
<a href="https://deeplearning4j.org/">Deeplearning4j</a>,
<a href="https://github.com/NVIDIA/caffe">NVIDIA Caffe</a>,
<a href="https://github.com/chainer/chainer">Chainer</a>,
<a href="https://github.com/PaddlePaddle/Paddle">PaddlePaddle</a>
					</p>
					<small><a href="https://en.wikipedia.org/wiki/List_of_datasets_for_machine_learning_research">List of datasets for machine learning research</a></small>
				</section>


				<section class="present" data-index-h="35" style="top: -240px; display: block;">
							<h1>Questions,  Thank you.</h1>
				</section>


			</div>
		</div>

		<script src="reveal.js/dist/reveal.js"></script>
		<script src="reveal.js/plugin/zoom/zoom.js"></script>
		<script src="reveal.js/plugin/notes/notes.js"></script>
		<script src="reveal.js/plugin/search/search.js"></script>
		<script src="reveal.js/plugin/markdown/markdown.js"></script>
		<script src="reveal.js/plugin/highlight/highlight.js"></script>
		<script>

			// Also available as an ES module, see:
			// https://revealjs.com/initialization/
			Reveal.initialize({
				controls: true,
				progress: true,
				center: true,
				hash: true,

				// Learn about plugins: https://revealjs.com/plugins/
				plugins: [ RevealZoom, RevealNotes, RevealSearch, RevealMarkdown, RevealHighlight ]
			});

		</script>

	</body>
</html>