PageSourceSearch

https://xiaoyizhang.me/

html xiaoyizhang.me collected 2026-09-29 15:28:40 UTC 13,822 bytes, 389 lines download raw bytes

1<!DOCTYPE html>
2<html lang="en-US">
3  <head>
4    <title>Xiaoyi Zhang | A\</title>
5
6    <meta charset="UTF-8">
7    <meta http-equiv="X-UA-Compatible" content="IE=edge">
8    <meta name="viewport" content="width=device-width, initial-scale=1">
9    <meta name="description" content="The Minimal Light is a simple and elegant jekyll theme for academic personal homepage.">
10    
11    <meta name="keywords" content="minimal light">
12    
13    
14    <link rel="canonical" href="https://minimal-light-theme.yliu.me/"/>
15    
16
17    <link rel="icon" media="(prefers-color-scheme:dark)" href="./assets/img/favicon-dark.png" type="image/png" />
18    <link rel="icon" media="(prefers-color-scheme:light)" href="./assets/img/favicon.png" type="image/png" />
19    
19<script src="./assets/js/favicon-switcher.js" type="application/javascript"></script>
19
20
21    <link rel=stylesheet href=https://cdnjs.cloudflare.com/ajax/libs/academicons/1.8.6/css/academicons.min.css integrity="sha256-uFVgMKfistnJAfoCUQigIl+JfUaP47GrRKjf6CTPVmw=" crossorigin=anonymous>
22    <link rel=stylesheet href=https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.11.2/css/all.min.css integrity="sha256-+N4/V/SbAFiW1MPBCXnfnP9QSN3+Keu+NlB+0ev/YKQ=" crossorigin=anonymous>
23
24    
25    <link rel="stylesheet" href="./assets/css/font_sans_serif.css">
26    
27    
28    <link rel="stylesheet" href="./assets/css/style.css">
29    <link rel="stylesheet" href="./assets/css/publications.css">
30    
31
32  </head>
33  <body>
34    <div class="wrapper">
35      <header>
36        
37        
38        <a class="image avatar"><img src="./assets/img/avatar.jpg" alt="avatar" /></a>
39        
40
41        
42        <a class="image avatar"><img src="./assets/img/nvshen.jpg" alt="avatar" /></a>
43        
44
45        <h1>Xiaoyi Zhang</h1>
46
47        
48        <position style="font-size:1.10rem;">Member of Technical Staff</position>
49        <br>
50        
51        
52        <autocolor>A\</autocolor>
53        <br>
54        <br>
55        
56        
57        <email>xiaoyizuw at gmail.com</email>
58        
59
60        <br>
61        <br>
62        <div class="social-icons">
63        
64        <a style="margin: 0 5px 0 0" href="https://scholar.google.com/citations?user=8qtpdWAAAAAJ">
65          <i class="ai ai-google-scholar" style="font-size:1.2rem"></i>
66        </a>  
67        
68
69        
70        <a style="margin: 0 5px 0 0" href="assets/files/CV_20240617.pdf">
71          <i class="ai ai-cv" style="font-size:1.3rem;"></i>
72        </a>
73        
74
75        
76
77        
78        <a style="margin: 0 5px 0 0" href="https://www.linkedin.com/in/xiaoyiz/">
79          <i class="fab fa-linkedin"></i>
80        </a>
81        
82
83        
84        <a style="margin: 0 0 0 0" href="https://x.com/xiaoyiz_uw">
85          <i class="fab fa-twitter"></i>
86        </a>
87        
88        </div>
89        <br>
90
91      </header>
92      <section>
93
94      <h2 id="about-me">About Me</h2>
95
96<p>I am currently at A\, focusing on HCI+AI research and product.</p>
97
98<p>Before this, I was a <strong>Staff Researcher &amp; Manager</strong> at Apple (2018-2024), passionate about building <strong>user‐friendly developer tools for ML engineers</strong> to boost their productivity. My team built ML infra and tools for model analysis, optimization, and testing, which is mission-critical for the launch of <strong>Apple Intelligence</strong> (<a href="https://machinelearning.apple.com/research/introducing-apple-foundation-models">Research post, Optimization section</a>) and <strong>Vision Pro</strong> (<a href="https://machinelearning.apple.com/research/talaria">Paper</a>).</p>
99
100<p>I received CS (HCI focus) PhD from UW, and interned at Google Research (with Kayur Patel on Colab) and MSR (with Meredith Ringel Morris). My first project at Apple was extending my PhD research to an ML-powered <a href="https://techcrunch.com/2020/12/03/iphones-can-now-automatically-recognize-and-label-buttons-and-ui-features-for-blind-users/">iOS accessibility feature</a> that supports 20M+ blind users’ daily life.</p>
101
102<p>Research Interests: <strong>Human-Computer Interaction (HCI)</strong>, ML Developer Tools, ML Model Optimization, and UI Accessibility.</p>
103
104<p><br /></p>
105
106<h2 id="achievements">Achievements:</h2>
107<ul>
108  <li><strong>Auto-OPT:</strong> ML infra to optimize all Apple Intelligence LLMs (on-device &amp; server).</li>
109  <li><strong>Talaria:</strong> Developer tool to enable Vision Pro ML models to run in low latency.</li>
110  <li><strong>Screen Recognition:</strong> The first on-device ML model to recognize and announce iOS App UI, benefiting 20M+ blind users.</li>
111  <li><strong>2 Best Papers, 3 Best Paper nominees, 5 Patents.</strong></li>
112</ul>
113
114<p><br /></p>
115<h2 id="publications" style="margin: 2px 0px -15px;">Selected Publications</h2>
116
117<div class="publications">
118Please see <b>full publication list (25+)</b> in <a href="https://scholar.google.com/citations?user=8qtpdWAAAAAJ">Google Scholar</a>
119<br /><br />
120
121<ol class="bibliography">
122
123
124
125<li>
126<div class="pub-row">
127  <div class="col-sm-9" style="position: relative;">
128      <div class="title"><a href="https://arxiv.org/pdf/2404.03085">Talaria: Interactively Optimizing Machine Learning Models for Efficient Inference</a></div>
129      <div class="author">Fred Hohman, Chaoqun Wang, Jinmook Lee, Jochen Görtler, Dominik Moritz, Jeffrey P. Bigham, Zhile Ren, Cecile Foret, Qi Shan, <strong>Xiaoyi Zhang</strong></div>
130      <div class="periodical"><em>CHI 2024</em>
131      </div>
132    <div class="links">
133       
134      <a href="https://arxiv.org/pdf/2404.03085" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
135      
136      
137      
138       
139      <a href="https://machinelearning.apple.com/research/talaria" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Video</a>
140      
141      
142       
143      <strong style="color:#2ea143; font-size: small">🏅 Best Paper Nominee; Critical for Apple Intelligence and Vision Pro models</strong>
144      
145      
146    </div>
147  </div>
148</div>
149</li>
150
151<br />
152
153
154
155<li>
156<div class="pub-row">
157  <div class="col-sm-9" style="position: relative;">
158      <div class="title"><a href="https://docs-assets.developer.apple.com/ml-research/papers/icon-labelling-mobile-apps-chi-22.pdf">Towards Complete Icon Labeling in Mobile Applications</a></div>
159      <div class="author">Jieshan Chen, Amanda Swearngin, Jason Wu, Titus Barik, Jeffrey Nichols, <strong>Xiaoyi Zhang</strong></div>
160      <div class="periodical"><em>CHI 2022</em>
161      </div>
162    <div class="links">
163       
164      <a href="https://docs-assets.developer.apple.com/ml-research/papers/icon-labelling-mobile-apps-chi-22.pdf" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
165      
166      
167       
168      <a href="https://machinelearning.apple.com/research/icon-labelling" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Project Page</a>
169      
170      
171      
172       
173      <strong style="color:#2ea143; font-size: small">Available in iOS</strong>
174      
175      
176    </div>
177  </div>
178</div>
179</li>
180
181<br />
182
183
184
185<li>
186<div class="pub-row">
187  <div class="col-sm-9" style="position: relative;">
188      <div class="title"><a href="https://docs-assets.developer.apple.com/ml-research/papers/screen-recognition-chi-2021.pdf">Screen Recognition: Creating Accessibility Metadata for Mobile Apps from Pixels</a></div>
189      <div class="author"><strong>Xiaoyi Zhang</strong>, Lilian de Greef, Amanda Swearngin, Samuel White, Kyle Murray, Lisa Yu, Qi Shan, Jeffrey Nichols, Jason Wu, Chris Fleizach, Aaron Everitt, Jeffrey P Bigham</div>
190      <div class="periodical"><em>CHI 2021</em>
191      </div>
192    <div class="links">
193       
194      <a href="https://docs-assets.developer.apple.com/ml-research/papers/screen-recognition-chi-2021.pdf" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
195      
196      
197       
198      <a href="https://machinelearning.apple.com/research/creating-accessibility-metadata" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Project Page</a>
199      
200       
201      <a href="https://www.youtube.com/watch?v=Z0pv0ZqFnC0" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Video</a>
202      
203      
204       
205      <strong style="color:#2ea143; font-size: small">🏆 Best Paper; Available in iOS</strong>
206      
207      
208    </div>
209  </div>
210</div>
211</li>
212
213<br />
214
215
216
217<li>
218<div class="pub-row">
219  <div class="col-sm-9" style="position: relative;">
220      <div class="title"><a href="https://arxiv.org/pdf/2105.01734">When Can Accessibility Help? Exploration of Accessibility Feature Recommendation on Mobile Devices</a></div>
221      <div class="author">Jason Wu, Gabriel Reyes, Sam C White, <strong>Xiaoyi Zhang</strong>, Jeffrey P Bigham</div>
222      <div class="periodical"><em>W4A 2021</em>
223      </div>
224    <div class="links">
225       
226      <a href="https://arxiv.org/pdf/2105.01734" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
227      
228      
229       
230      <a href="https://machinelearning.apple.com/research/recommending-accessibility" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Project Page</a>
231      
232      
233      
234       
235      <strong style="color:#2ea143; font-size: small">🏆 Best Paper</strong>
236      
237      
238    </div>
239  </div>
240</div>
241</li>
242
243<br />
244
245
246
247<li>
248<div class="pub-row">
249  <div class="col-sm-9" style="position: relative;">
250      <div class="title"><a href="https://faculty.washington.edu/wobbrock/pubs/assets-18.01.pdf">Examining Image-Based Button Labeling for Accessibility in Apps through Large-Scale Analysis</a></div>
251      <div class="author">Anne Spencer Ross, <strong>Xiaoyi Zhang</strong>, James Fogarty, Jacob O Wobbrock</div>
252      <div class="periodical"><em>ASSETS 2018</em>
253      </div>
254    <div class="links">
255       
256      <a href="https://faculty.washington.edu/wobbrock/pubs/assets-18.01.pdf" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
257      
258      
259      
260      
261      
262       
263      <strong style="color:#2ea143; font-size: small">🏅 Best Paper Nominee</strong>
264      
265      
266    </div>
267  </div>
268</div>
269</li>
270
271<br />
272
273
274
275<li>
276<div class="pub-row">
277  <div class="col-sm-9" style="position: relative;">
278      <div class="title"><a href="https://faculty.washington.edu/wobbrock/pubs/assets-17.pdf">Epidemiology as a Framework for Large-scale Mobile Application Accessibility Assessment</a></div>
279      <div class="author">Anne Spencer Ross, <strong>Xiaoyi Zhang</strong>, James Fogarty, Jacob O Wobbrock</div>
280      <div class="periodical"><em>ASSETS 2017</em>
281      </div>
282    <div class="links">
283       
284      <a href="https://faculty.washington.edu/wobbrock/pubs/assets-17.pdf" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
285      
286      
287      
288      
289      
290       
291      <strong style="color:#2ea143; font-size: small">🏅 Best Paper Nominee</strong>
292      
293      
294    </div>
295  </div>
296</div>
297</li>
298
299<br />
300
301
302
303<li>
304<div class="pub-row">
305  <div class="col-sm-9" style="position: relative;">
306      <div class="title"><a href="https://faculty.washington.edu/wobbrock/pubs/chi-17.03.pdf">Interaction Proxies for Runtime Repair and Enhancement of Mobile Application Accessibility</a></div>
307      <div class="author"><strong>Xiaoyi Zhang</strong>, Anne Spencer Ross, Anat Caspi, James Fogarty, Jacob O Wobbrock</div>
308      <div class="periodical"><em>CHI 2017</em>
309      </div>
310    <div class="links">
311       
312      <a href="https://faculty.washington.edu/wobbrock/pubs/chi-17.03.pdf" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
313      
314      
315      
316       
317      <a href="https://www.youtube.com/watch?v=v15FfDAsf04" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Video</a>
318      
319      
320       
321      <strong style="color:#2ea143; font-size: small">Foundation Work for Screen Recognition</strong>
322      
323      
324    </div>
325  </div>
326</div>
327</li>
328
329<br />
330
331
332
333<li>
334<div class="pub-row">
335  <div class="col-sm-9" style="position: relative;">
336      <div class="title"><a href="https://www.microsoft.com/en-us/research/uploads/prod/2017/01/gazespeakUpdated.pdf">Smartphone-Based Eye Gesture Communication for People with Motor Disabilities</a></div>
337      <div class="author"><strong>Xiaoyi Zhang</strong>, Harish Kulkarni, Meredith Ringel Morris</div>
338      <div class="periodical"><em>CHI 2017</em>
339      </div>
340    <div class="links">
341       
342      <a href="https://www.microsoft.com/en-us/research/uploads/prod/2017/01/gazespeakUpdated.pdf" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">PDF</a>
343      
344      
345      
346       
347      <a href="https://www.youtube.com/watch?v=Nffiaaw9PvQ" class="btn btn-sm z-depth-0" role="button" target="_blank" style="font-size:12px;">Video</a>
348      
349      
350       
351      <strong style="color:#2ea143; font-size: small">Patent <a href="https://patents.google.com/patent/US10353475B2/en" target="_blank">(Link)</a></strong>
352      
353      
354    </div>
355  </div>
356</div>
357</li>
358
359<br />
360
361
362
363</ol>
364</div>
365
366
367
368      <br>
369
370      
371
372      </section>
373      <footer>
374        
375      </footer>
376    </div>
377    
377<script src="/assets/js/scale.fix.js"></script>
377
378    
379    
379<script>
vendor: 364 bytes, lines 379-384
379
380      (function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
381      (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),
382      m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
383      })(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
384      ga('create', '
384UA-111540567-4
vendor: 46 bytes, lines 384-386
384', 'auto');
385      ga('send', 'pageview');
386    
386</script>
386
387    
388  </body>
389</html>

Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.