-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
386 lines (351 loc) · 20.5 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
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
---
title: Ziwei Gu
extra_css: "mystyle.css"
---
<div class="container">
<div class="row">
<div class="col-md-3">
<img style="width:100%;max-width:100%;border-radius:25px;margin-top:10%" src="/assets/images/ziwei.jpg" alt="profile photo">
<h1>Ziwei Gu</h1>
<!-- <p style="text-align:center;">
PhD Candidate, Computer Science, Stanford University
</p> -->
<br>
<div class="text-center">
<a style="padding-right:20px !important" href="mailto:[email protected]"><i class="fa fa-envelope-o fa-2x"></i></a>
<a style="padding-right:20px !important" href="/assets/data/cv.pdf"><i class="ai ai-cv ai-2x"></i></a>
<a style="padding-right:20px !important" href="https://scholar.google.com/citations?user=w_PbwxEAAAAJ&hl=en"><i class="ai ai-google-scholar-square ai-2x"></i></a>
<a style="padding-right:20px !important" href="https://github.com/ZiweiGu/"><i class="fa fa-github fa-2x"></i></a>
<a href="https://twitter.com/ziweigu"><i class="fa fa-twitter fa-2x"></i></a>
</div>
<br>
<br>
<h3 style="text-align:center;">Updates</h3>
<p>
<strong>03/2024</strong> - Our eye-tracking paper was accepted at CHI'24 Late Breaking Work.
</p>
<p>
<strong>01/2024</strong> - Two papers conditionally accepted at CHI'24. Congratulations to my co-authors!
</p>
<!-- <p>
<strong>05/2021</strong> - I graduated from Cornell!
</p>
<p>
<strong>01/2021</strong> - My first first-author paper on <a href="/assets/data/sensemaking.pdf">user sensemaking</a> was accepted at WWW'21.
</p>
<p>
<strong>01/2021</strong> - Our paper on <a href="/assets/data/tessera.pdf">mining interaction logs</a> was accepted at CHI'21.
</p>
<p>
<strong>06/2020</strong> - I started an internship at Lyft as a data scientist intern.
</p>
<p>
<strong>01/2020</strong> - Our paper on <a href="/assets/data/silva.pdf">Silva</a> was accepted at CHI'20. Congratulations to my co-authors!
</p> -->
</div>
<div style="margin-left:30px" class="col-md-8">
<table style="width:100%;max-width:800px;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tbody>
<tr style="padding:0px">
<td style="padding:0px">
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:3%;width:63%;vertical-align:middle">
<!-- <p style="text-align:center">
<name>Ziwei Gu</name>
</p> -->
<h5>Hi, I'm Ziwei.</h5>
<p>I am a second-year Ph.D. student in Computer Science at <a href="https://www.harvard.edu/">Harvard University</a>, advised by <a href="https://glassmanlab.seas.harvard.edu/">Dr. Elena Glassman</a>.
<!-- My research interests lie at the intersection of human-computer interaction and natural language processing. -->
Currently, my research focuses on <i>augmenting human cognition and efficiency</i> by leveraging large language models (LLMs) and interactive techniques, aiming to <i>ensure that potential AI errors can be easily noticed, judged, and recovered from</i>, a concept I formalize as <strong>AI-resiliency</strong>. <!-- My current research focuses on developing <strong>novel interactive systems</strong> to (1) help people <strong>perceive patterns in large language corpora</strong> and (2) <strong>support human-centered AI</strong>. -->
</p>
<p>
Before coming to Harvard, I graduated with a bachelors in Mathematics and Computer Science (December 2020) and masters in Computer Science (May 2021) from <a href="https://www.cornell.edu/">Cornell University</a>.
</p>
<!-- <p style="color:#006600;">I am actively seeking PhD advisors for Fall 2022.</p> -->
<!-- <p style="text-align:center">
<a href="mailto:[email protected]">Email</a>  / 
<a href="/assets/data/cv.pdf">CV</a>  / 
<a href="/assets/data/transcript.pdf">Transcript</a>  / 
<a href="https://scholar.google.com/citations?user=w_PbwxEAAAAJ&hl=en">Google Scholar</a>  / 
<a href="https://twitter.com/ziweigu">Twitter</a>  / 
<a href="https://github.com/ZiweiGu/">Github</a>
</p> -->
</td>
</tr>
</tbody></table>
<!-- <table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<heading>Research</heading>
<p>
I'm interested in human-computer interaction, machine learning, data mining, and algorithm fairness. My current research focuses on developing <strong>novel interactive systems</strong> to help people <strong>make sense of data and fairness issues</strong>.
</p>
</td>
</tr>
</tbody></table> -->
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<heading>Publications</heading>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two">
<img src='/assets/images/gptsm_eyetracking.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="/assets/data/gptsm_eyetracking.pdf">
<papertitle>Why Do Skimmers Perform Better with Grammar-Preserving Text Saliency Modulation (GP-TSM)? Evidence from an Eye Tracking Study</papertitle>
</a>
<br>
<strong>Ziwei Gu</strong>,
<a href="https://scholar.google.com/citations?user=pgMGUd4AAAAJ&hl=en">Owen Raymond</a>,
<a href="https://cs.colby.edu/nsalmadi/">Naser Al Madi</a>,
<a href="https://glassmanlab.seas.harvard.edu/glassman.html">Elena L. Glassman</a>
<br>
<em>CHI</em> 2024 Late Breaking Work
<br>
<a href="/assets/data/gptsm_eyetracking.pdf">paper</a>
/
<a href="https://drive.google.com/file/d/1b_sx_DodP5IRwv5k20XpOg04N9HKEV11/view?usp=sharing">video</a>
<!-- /
<a href="/">slides</a> -->
<p></p>
<p>How can we get a better understanding of the mechanism through which an LLM-based reading assistance tool (GP-TSM) supports reading? We conducted an eye-tracking user study with 24 participants, followed by an analysis of the unique gaze patterns associated with GP-TSM. </p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two">
<img src='/assets/images/gptsm.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="/assets/data/gptsm.pdf">
<papertitle>An AI-Resilient Text Rendering Technique for Reading and Skimming Documents</papertitle>
</a>
<br>
<strong>Ziwei Gu</strong>,
<a href="https://ianarawjo.com/">Ian Arawjo</a>,
<a href="https://likenneth.github.io/">Kenneth Li</a>,
<a href="https://jkk.name/">Jonathan K. Kummerfeld</a>,
<a href="https://glassmanlab.seas.harvard.edu/glassman.html">Elena L. Glassman</a>
<br>
<em>CHI</em> 2024
<br>
<a href="/assets/data/gptsm.pdf">paper</a>
/
<a href="https://drive.google.com/file/d/1ZF0wOoBp8FezZtGN5JXP5_Oo9r999YVD/view?usp=sharing">video</a>
<!-- /
<a href="/">slides</a> -->
<p></p>
<p>We propose the idea of "AI-resilience" and an LLM-powered technique that supports reading through recursive summarization while allowing readers to easily notice and recover from LLM summaries they disagree with.</p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two">
<img src='/assets/images/llm_sensemaking.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="/assets/data/llm_sensemaking.pdf">
<papertitle>Supporting Sensemaking of Large Language Model Outputs at Scale</papertitle>
</a>
<br>
<a href="https://www.katygero.com/">Katy Ilonka Gero</a>,
<a href="https://seas.harvard.edu/person/chelse-swoopes">Chelse Swoopes</a>,
<strong>Ziwei Gu</strong>,
<a href="https://jkk.name/">Jonathan K. Kummerfeld</a>,
<a href="https://glassmanlab.seas.harvard.edu/glassman.html">Elena L. Glassman</a>
<br>
<em>CHI</em> 2024
<br><span style="background-color: hwb(45 70% 1%);">Honorable Mention Award</span>
<br>
<a href="/assets/data/llm_sensemaking.pdf">paper</a>
<!-- /
<a href="/">video</a>
/
<a href="/">slides</a> -->
<p></p>
<p>Large language models (LLMs) are capable of generating multiple responses to a single prompt, yet little effort has been expended to help people make use of this capability. In this paper, we explore how to present many LLM responses at once.</p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='tessera_image'>
<img src='/assets/images/tessera.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://dl.acm.org/doi/pdf/10.1145/3411764.3445728/">
<papertitle>Tessera: Discretizing Data Analysis Workflows on a Task Level</papertitle>
</a>
<br>
<a href="https://nathanyanjing.github.io/">Jing Nathan Yan</a>,
<strong>Ziwei Gu</strong>,
<a href="https://jeffrz.com/">Jeffrey M Rzeszotarski</a>
<br>
<em>CHI</em> 2021
<br>
<a href="/assets/data/tessera.pdf">paper</a>
/
<a href="https://www.youtube.com/watch?v=JAhsVyyURBw">video</a>
/
<a href="https://docs.google.com/presentation/d/1Oyme81Ke2jHWY6lAFPXxAWY6xqac-kWe/edit?usp=sharing&ouid=112404362805911807942&rtpof=true&sd=true">slides</a>
<p></p>
<p>Interaction logs can be extremely complex yet useful. Breaking down event logs into goal-directed segments can make it easier to understand user workflow.</p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='sensemaking_image'>
<img src='/assets/images/sensemaking.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://dl.acm.org/doi/pdf/10.1145/3442381.3450092/">
<papertitle>Understanding User Sensemaking in Machine Learning Fairness Assessment Systems</papertitle>
</a>
<br>
<strong>Ziwei Gu</strong>,
<a href="https://nathanyanjing.github.io/">Jing Nathan Yan</a>,
<a href="https://jeffrz.com/">Jeffrey M Rzeszotarski</a>
<br>
<em>WWW</em> 2021
<br>
<a href="/assets/data/sensemaking.pdf">paper</a>
/
<a href="https://www.youtube.com/watch?v=AlXJe8NbWdQ">video</a>
/
<a href="https://docs.google.com/presentation/d/1zts6eChTxsgiXDT4vsmLT5vAfNra_Zw7/edit?usp=sharing&ouid=112404362805911807942&rtpof=true&sd=true">slides</a>
<p></p>
<p>We ask a fundamental research question: How do core design elements of debiasing systems shape how people reason about biases? We present distinctive sensemaking patterns and surprising findings from think-aloud studies.</p>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='silva_image'>
<img src='/assets/images/silva.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://dl.acm.org/doi/pdf/10.1145/3313831.3376447/">
<papertitle>Silva: Interactively Assessing Machine Learning Fairness Using Causality</papertitle>
</a>
<br>
<a href="https://nathanyanjing.github.io/">Jing Nathan Yan</a>,
<strong>Ziwei Gu</strong>,
<a href="https://www.cs.cornell.edu/~hubert/">Hubert Lin</a>,
<a href="https://jeffrz.com/">Jeffrey M Rzeszotarski</a>
<br>
<em>CHI</em> 2020
<br>
<a href="/assets/data/silva.pdf">paper</a>
/
<a href="https://www.youtube.com/watch?v=tNNibjB1RHw">short video</a>
/
<a href="https://www.youtube.com/watch?v=sbccxPIOUec">long video</a>
/
<a href="https://docs.google.com/presentation/d/1IkqsQ2esSQu4gY6pxtZNWEymRcGNP6h_/edit?usp=sharing&ouid=112404362805911807942&rtpof=true&sd=true">slides</a>
<p></p>
<p>We present Silva, an interactive tool that utilizes a causal graph linked with quantitative metrics to help people find and reason about sources of biases in datasets and machine learning models.</p>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<heading>Technical Reports</heading>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<div class="one">
<div class="two" id='noie_image'>
<img src='/assets/images/noie.png' width="160"></div>
</div>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://www.wesg.me/publication/noie/NOIE.pdf">
<papertitle>Neural Open Information Extraction with Transformers</papertitle>
</a>
<br>
<a href="https://www.wesg.me/">Wes Gurnee</a>,
<strong>Ziwei Gu</strong>
<br>
<br>
<a href="/assets/data/noie.pdf">paper</a>
/
<a href="https://github.com/wesg52/NOIE">code</a>
/
<a href="https://docs.google.com/presentation/d/1yjNsdxP_916y4ioL78wqaqaOmmfMMCZSOI35QM5oWVo/edit?usp=sharing">slides</a>
<p></p>
<p>We trained a deep learning model (Transformer) for open information extraction, modeled as a sequence to sequence transduction task. We showed that our model was competitive with state-of-the-art systems, but without the dependencies on other NLP tools.</p>
</td>
</tr>
</tbody></table>
<!--
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20"><tbody>
<tr>
<td>
<heading>Service</heading>
</td>
</tr>
</tbody></table> -->
<!-- <table width="100%" align="center" border="0" cellpadding="20"><tbody>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle"><img src="images/cvf.jpg"></td>
<td width="75%" valign="center">
<a href="http://cvpr2021.thecvf.com/area-chairs">Area Chair, Longuet-Higgins Award Committee Member, CVPR 2021</a>
<br><br>
<a href="http://cvpr2019.thecvf.com/area_chairs">Area Chair, CVPR 2019</a>
<br><br>
<a href="http://cvpr2018.thecvf.com/organizers/area_chairs">Area Chair, CVPR 2018</a>
</td>
</tr>
<tr>
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/cs188.jpg" alt="cs188">
</td>
<td width="75%" valign="center">
<a href="http://inst.eecs.berkeley.edu/~cs188/sp11/announcements.html">Graduate Student Instructor, CS188 Spring 2011</a>
<br>
<br>
<a href="http://inst.eecs.berkeley.edu/~cs188/fa10/announcements.html">Graduate Student Instructor, CS188 Fall 2010</a>
<br>
<br>
<a href="http://aima.cs.berkeley.edu/">Figures, "Artificial Intelligence: A Modern Approach", 3rd Edition</a>
</td>
</tr>
</tbody></table> -->
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:0px">
<br>
<p style="text-align:center;font-size:small;">
Layout inspired by <a href="https://github.com/jonbarron/jonbarron_website">this template</a>
</p>
</td>
</tr>
</tbody></table>
</td>
</tr>
</tbody>
</table>
</div>
</div>
</div>