-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
336 lines (282 loc) · 16.8 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
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
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- Meta tags for social media banners, these should be filled in appropriatly as they are your "business card" -->
<!-- Replace the content tag with appropriate information -->
<meta name="description" content="DESCRIPTION META TAG">
<meta property="og:title" content="SOCIAL MEDIA TITLE TAG"/>
<meta property="og:description" content="SOCIAL MEDIA DESCRIPTION TAG TAG"/>
<meta property="og:url" content="URL OF THE WEBSITE"/>
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X630-->
<meta property="og:image" content="static/image/your_banner_image.png" />
<meta property="og:image:width" content="1200"/>
<meta property="og:image:height" content="630"/>
<meta name="twitter:title" content="TWITTER BANNER TITLE META TAG">
<meta name="twitter:description" content="TWITTER BANNER DESCRIPTION META TAG">
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X600-->
<meta name="twitter:image" content="static/images/your_twitter_banner_image.png">
<meta name="twitter:card" content="summary_large_image">
<!-- Keywords for your paper to be indexed by-->
<meta name="keywords" content="KEYWORDS SHOULD BE PLACED HERE">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Academic Project Page</title>
<link rel="icon" type="image/x-icon" href="static/images/favicon.ico">
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="static/css/bulma.min.css">
<link rel="stylesheet" href="static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="static/css/bulma-slider.min.css">
<link rel="stylesheet" href="static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="static/css/index.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="https://documentcloud.adobe.com/view-sdk/main.js"></script>
<script defer src="static/js/fontawesome.all.min.js"></script>
<script src="static/js/bulma-carousel.min.js"></script>
<script src="static/js/bulma-slider.min.js"></script>
<script src="static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title"><span class="b1">PrivacyOracle</span> : Configuring Sensor Privacy
Firewalls with Large Language Models in Smart
Built Environments</h1>
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://scholar.google.com/citations?user=ote_P0QAAAAJ&hl=en&oi=ao" target="_blank">Brian Wang</a><sup>1</sup>,</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=F6Gzg9gAAAAJ&hl=en&oi=sra" target="_blank">Luis Antonio Garcia</a><sup>2</sup>,</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=X2Qs7XYAAAAJ&hl=en&oi=ao" target="_blank">Mani Srivastava</a><sup>1</sup>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup> University of California Los Angeles, <sup>2</sup> University of Utah <br>SafeThings 2024</span>
<!--<span class="eql-cntrb"><small><br><sup>*</sup>Indicates Equal Contribution</small></span>-->
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- Arxiv PDF link -->
<span class="link-block">
<a href="https://ieeexplore.ieee.org/abstract/document/10579527" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span>
<!-- Supplementary PDF link -->
<span class="link-block">
<a href="static/pdfs/poster.pdf" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Poster</span>
</a>
</span>
<!-- Github link -->
<span class="link-block">
<a href="https://github.com/nesl/PrivacyOracle" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
<!-- ArXiv abstract Link -->
<!-- <span class="link-block">
<a href="https://arxiv.org/abs/<ARXIV PAPER ID>" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span> -->
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Paper Section -->
<section class="section hero is-light">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Abstract</h2>
<div class="content has-text-justified">
<p>
Modern smart buildings and environments rely on sensory infrastructure to capture and process information about their inhabitants. However, it remains challenging to ensure that this infrastructure complies with privacy norms, preferences, and regulations; individuals occupying smart environments are often occupied with their tasks, lack awareness of the surrounding sensing mechanisms, and are non-technical experts. This problem is only exacerbated by the increasing number of sensors being deployed in these environments, as well as services seeking to use their sensory data. As a result, individuals face an unmanageable number of privacy decisions, preventing them from effectively behaving as their own “privacy firewall” for filtering and managing the multitude of personal information flows. These decisions often require qualitative reasoning over privacy regulations, understanding privacy-sensitive contexts, and applying various privacy transformations when necessary We propose the use of Large Language Models (LLMs), which have demonstrated qualitative reasoning over social/legal norms, sensory data, and program synthesis, all of which are necessary for privacy firewalls. We present PrivacyOracle, a prototype system for configuring privacy firewalls on behalf of users using LLMs, enabling automated privacy decisions in smart built environments. Our evaluation shows that PrivacyOracle achieves up to 98% accuracy in identifying privacy-sensitive states from sensor data, and demonstrates 75% accuracy in measuring social acceptability of information flows.
</p>
</div>
</div>
</div>
</div>
</section>
<!-- End paper abstract -->
<!-- Image carousel -->
<!--
<section class="hero is-small">
<div class="hero-body">
<div class="container">
<div id="results-carousel" class="carousel results-carousel">
<div class="item">
<img src="static/images/carousel1.jpg" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
First image description.
</h2>
</div>
<div class="item">
<img src="static/images/carousel2.jpg" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Second image description.
</h2>
</div>
<div class="item">
<img src="static/images/carousel3.jpg" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Third image description.
</h2>
</div>
<div class="item">
<img src="static/images/carousel4.jpg" alt="MY ALT TEXT"/>
<h2 class="subtitle has-text-centered">
Fourth image description.
</h2>
</div>
</div>
</div>
</div>
</section> -->
<!-- End image carousel -->
<section class="hero is-small b2">
<div class="hero-body">
<div class="container">
<div class="columns is-centered has-text-centered">
<h2 class="title is-3">Why a privacy firewall, and what is it?</h2>
</div>
<div class="content has-text-justified">
We face novel privacy risks in today's smart environments, where overcollection and invasive inferences over personal sensory data have become the norm. As the number of physical sensing devices and applications executed in such environments continue to grow, personal privacy will only become more difficult to manage. In order to tackle the issues of manging personal privacy in unregulated sensing environments, we introduce the idea of <span class="b1">privacy firewalls</span>. Privacy firewalls mediate the flow of information bewteen sensing infrastructure and service providers, and do so automatically on behalf of data subjects whose data is being collected.
</div>
<div class="columns is-centered has-text-centered">
<img src="static/images/motivation.png" alt="MY ALT TEXT" class="b4"/>
</div>
<h2 class="subtitle has-text-centered">
High level overview of privacy firewalls
</h2>
<div class="content has-text-justified">
Privacy firewalls must possess several capabilities: Firstly, qualititative reasoning about data-sharing decisions which necessarily involve legal rules, social values, utility of services, and personal privacy preferences. Secondly, they must infer different states of users in the environment, and decide whether a particular state of a user (e.g. in a bathroom) presents privacy risks. Lastly, a privacy firewall should be able to quickly adapt and use tools for preserving privacy under different privacy/utility tradeoffs.
</div>
</div>
</div>
</section>
<section class="hero is-small b3 is-light">
<div class="hero-body">
<div class="container">
<div class="columns is-centered has-text-centered">
<h2 class="title is-3">Overview of PrivacyOracle</h2>
</div>
<div class="content has-text-justified">
PrivacyOracle is a Large Language Model (LLM) based privacy firewall. We achieve each of the aforementioned tasks using the informational flow verification, sensitive state detection, and tool selection modules of our system. Each module has a prompt template which interacts with a LLM.
</div>
<div class="columns is-centered has-text-centered">
<img src="static/images/system.png" alt="MY ALT TEXT" class="b5"/>
</div>
<h2 class="subtitle has-text-centered">
Architecture of PrivacyOracle, our privacy firewall for regulating the flow of sensory data in smart-built environments via several LLM tasks.
</h2>
<div class="content has-text-justified">
The first service is to verify information flow requests from a service provider to a sensing infrastructure. This service must identify the necessary knowledge to configure the privacy rules for sharing these informational flows. We use the same definition of informational flows as described in the theory of Contextual Integrity (CI).
The second service in a privacy firewall is to identify sensitive segments of sensor data that should be hidden based on privacy preferences. This service requires sensor data to be augmented with metadata information (such as sensor type, location, and names), that grants additional context for identifying the sensitivity of data. This task filters out sensor data based on a privacy preference expressed in natural language.
The last service of a privacy firewall is to identify the appropriate tools for transforming sensor data into a format that is acceptable from a privacy perspective. This allows privacy firewalls to not only accept or reject informational flows but modify them such that they are acceptable. Tool selection involves reasoning over both the privacy preferences of users and a library of tools, which allows automated selection of a particular tool given different scenarios. Furthermore, this service not only selects tools but can generate dataflow pipelines using other preprocessing tools as well, which improves the interoperability of different tools without having to worry about modifying the tool interfaces.
</div>
</div>
</div>
</section>
<section class="hero is-small b7">
<div class="hero-body">
<div class="container">
<div class="columns is-centered has-text-centered">
<h2 class="title is-3">Results</h2>
</div>
<div class="content has-text-justified">
We evaluate our system on several privacy tasks, and use GPT-3.5 and GPT-4.0 as our LLM.
</div>
<!--Confusion matrix -->
<div class="columns is-centered has-text-centered">
<img src="static/images/cm.JPG" alt="MY ALT TEXT" class="b6"/>
</div>
<h2 class="subtitle has-text-centered lowpad">
Difference in acceptability scores of information flows with given recipient/transmission principle pairs, measured between LLM and a previous user study (see paper for reference)
</h2>
<!--Sensitive State Detection-->
<div class="columns is-centered has-text-centered">
<img src="static/images/sensitive_states.png" alt="MY ALT TEXT" class="b6 noborder"/>
</div>
<h2 class="subtitle has-text-centered lowpad">
Detection accuracy of GPT-4 for privacy-sensitive states from sensor data.
</h2>
<!--Pipeline example for tool selection-->
<div class="columns is-centered has-text-centered">
<img src="static/images/pipeline_example.png" alt="MY ALT TEXT" class="b4"/>
</div>
<h2 class="subtitle has-text-centered lowpad">
Response examples of GPT-4 for adapting privacy tools to natural language requirements
</h2>
<!--Privacy utility scores-->
<div class="columns is-centered has-text-centered">
<img src="static/images/tradeoffs.png" alt="MY ALT TEXT" class="b6 noborder"/>
</div>
<h2 class="subtitle has-text-centered lowpad">
Privacy (inferred age, gender, race) and utility scores (emotion) of pipelines generated by different privacy requirements (see paper/poster for more details).
</h2>
</div>
</div>
</section>
<!--BibTex citation -->
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@inproceedings{wang2024privacyoracle,
title={PrivacyOracle: Configuring Sensor Privacy Firewalls with Large Language Models in Smart Built Environments},
author={Wang, Brian and Garcia, Luis Antonio and Srivastava, Mani},
booktitle={2024 IEEE Security and Privacy Workshops (SPW)},
pages={239--245},
year={2024},
organization={IEEE}
}</code></pre>
</div>
</section>
<!--End BibTex citation -->
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This page was built using the <a href="https://github.com/eliahuhorwitz/Academic-project-page-template" target="_blank">Academic Project Page Template</a> which was adopted from the <a href="https://nerfies.github.io" target="_blank">Nerfies</a> project page.
You are free to borrow the of this website, we just ask that you link back to this page in the footer. <br> This website is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
<!-- Statcounter tracking code -->
<!-- You can add a tracker to track page visits by creating an account at statcounter.com -->
<!-- End of Statcounter Code -->
</body>
</html>