-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathindex.html
596 lines (572 loc) · 34.5 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
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<title>Hanshi Sun</title>
<link rel="shortcut icon" href="assets/img/title.ico">
<link rel="stylesheet" href="red.css" type="text/css">
<style>
@font-face {
font-family: 'TriForceFont';
src: url('assets/Triforce.ttf') format('truetype');
}
.TriForceFont {
font-family: 'TriForceFont', sans-serif !important;
font-size: 3.0rem;
}
</style>
</head>
<body>
<table>
<tbody>
<tr>
<td>
<img src="assets/img/bio.jpg" border="0" width="230"><br>(at Lake Louise, Alberta, Canada)
</td>
<td style="padding:0.5%;width:70%;vertical-align:middle">
<div id="toptitle">
<h1 style="color: black;">Hanshi Sun   孙寒石</h1>
</div>
<p>I am a M.S. student in ECE at <a href=https://www.cmu.edu>Carnegie Mellon University</a>. Now I am working on LLM efficiency with Prof. <a
href="https://www.andrew.cmu.edu/user/beidic/">Beidi Chen</a>. I also work closely with Prof. <a href="https://azanette.com/">Andrea Zanette</a>. I obtained my bachelor degree at <a
href="https://www.seu.edu.cn/english/" target="blank">Southeast University</a>. I was a
member of the <a href="http://palm.seu.edu.cn/members.html">PAttern Learning and Mining (PALM)
Lab</a> , where I am fortunate to be advised by Prof. <a
href="https://scholar.google.co.uk/citations?user=EnDCJKMAAAAJ&hl=en">Yi Zhou</a>. Before, I was
a research intern at the <a href=https://www.ualberta.ca/index.html target=_blank
rel=noopener>University of Alberta</a>, supervised by Prof. <a
href="https://www.ece.ualberta.ca/~xingyu/index.html">Xingyu Li</a>.
<br><br>
<img src="assets/img/gmail.png" width="20px" style="margin-bottom:-3px"> hanshis [at] andrew [dot] cmu [dot] edu<br><br>
</p>
<p style="line-height: 1.8;">
<a href="https://twitter.com/preminstrel">
<img src="assets/img/X.png" width="20px" style="margin-bottom:-3px"> @preminstrel
</a><br>
<a href="https://github.com/preminstrel">
<img src="assets/img/github.jpeg" width="20px" style="margin-bottom:-3px">
GitHub (preminstrel)
</a><br>
<a href="https://scholar.google.com/citations?user=BjQHEh8AAAAJ&hl=en">
<img src="assets/img/google-scholar.png" width="20px" style="margin-bottom:-3px">
Google Scholar
</a><br>
<a href="https://www.linkedin.com/in/hanshi-sun-5b74b8228/" target="blank">
<img src="assets/img/LinkedIn.png" width="20px" style="margin-bottom:-3px">
LinkedIn
</a><br>
<a href="https://preminstrel.github.io/blog" target="blank">
<img src="assets/img/blog.jpeg" width="20px" style="margin-bottom:-3px">
Blog
</a>
</p>
</td>
</tr>
<tr>
</tr>
</tbody>
</table>
<h2>News</h2>
<ul>
<div style="height:170px;width:fit-content;overflow:auto;background:#FFFFFF;">
<li>
<p style="line-height: 1.5;">[2024/10/01] Our <a href="https://Zanette-Labs.github.io/SpeculativeRejection" target="blank">Speculative Rejection</a> has been accepted by <a href="https://nips.cc/" target="blank"><img src="assets/img/NeurIPS logo.svg" width="28px" style="vertical-align:middle"> NeurIPS 2024</a>! See you in Vancouver!
</p>
</li>
<li>
<p style="line-height: 1.5;">[2024/07/09] Our <a href="https://infini-ai-lab.github.io/TriForce/" target="blank">TriForce</a> has been accepted by <a href="https://colmweb.org/" target="blank">🦙 COLM 2024</a>! See you in Philadelphia!
</p>
</li>
<li>
<p style="line-height: 1.5;">[2024/06/03] Work as a MLSys Research Intern in Seed-Foundation team at ByteDance (Seattle Office) <img src="assets/img/ByteDance Logo English.png" width="100px" style="vertical-align:middle" >.
</p>
</li>
<li>
<p style="line-height: 1.5;">[2023/11/06] Join Prof. <a
href="https://www.andrew.cmu.edu/user/beidic/">Beidi Chen</a>'s group at CMU <img src="assets/img/cmu_logo.jpeg" height="28px" style="vertical-align:middle"> <img src="assets/img/cmu_logo_2.png" height="26px" style="vertical-align:middle"></a>.
</p>
</li>
<li>
<p style="line-height: 1.5;">[2023/06/20] Graduate from <a
href="https://www.seu.edu.cn/english/" target="blank">Southeast University</a> <img src="assets/img/seu_l.jpeg" height="28px" style="vertical-align:middle"> with a bachelor degree <img src="assets/img/seu_ee.jpeg" height="28px" style="vertical-align:middle"> <img id="school_logo" src="assets/img/PALMLogo.png" height="28px" style="vertical-align:middle">.</p>
</li>
<li>
<p style="line-height: 1.5;">[2022/11/15] I am working at <img src="assets/img/apple.png" width=15px" style="vertical-align:middle"> Apple as an intern in EE/SW team for eight months.
</p>
</li>
<li>
<p style="line-height: 1.5;">[2022/07/06] I am working as a research intern during summer 2022 in Prof.<a
href=https://www.ece.ualberta.ca/~xingyu/index.html target=_blank rel=noopener>
Xingyu Li</a>’s group at the <a href=https://www.ualberta.ca/index.html target=_blank
rel=noopener>University of Alberta</a> <img src="assets/img/University_of_Alberta_logo.png" width="80px" style="vertical-align:middle">.
</p>
</li>
</div>
</ul>
<h2>Publications</h2>
<style>
.button {
border: none;
background-color: transparent;
padding: 0;
font: inherit;
text-decoration: underline;
cursor: pointer;
outline: none;
color: black;
}
.selected-button {
border: none;
background-color: transparent;
padding: 0;
font: inherit;
text-decoration: underline;
cursor: pointer;
outline: none;
color: #089797;
}
</style>
<p>
<button class="button selected-button" id="button1" onclick="showSelectedPublications(); selectButton('button1')">Representative Papers</button> /
<button class="button" id="button2" onclick="showFullPublications(); selectButton('button2')">Full Papers by Year</button>.
MLSys papers are <span style="background-color: #ffffd0;">highlighted</span>.
</p>
<div id="selected_publications">
<table style="background-color: #ffffd0;">
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/shadowkv_sys.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://github.com/bytedance/ShadowKV" target="blank"><b><img src="assets/img/ShadowKV.png" width="24px"> ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference</a>
</b> <br>
<br>
<strong>Hanshi Sun</strong>, <a href="https://lchang20.github.io/">Li-Wen Chang</a>, <a href="https://sites.google.com/view/wenleibao/">Wenlei Bao</a>, <a href="https://sizezheng.github.io/">Size Zheng</a>, <a href="https://zheng-ningxin.github.io/">Ningxin Zheng</a>, <a href="https://scholar.google.com/citations?user=ZMfk2F8AAAAJ&hl=zh-CN">Xin Liu</a>, <a href="https://www.andrew.cmu.edu/user/harryd/">Harry Dong</a>, <a href="https://users.ece.cmu.edu/~yuejiec/index.html">Yuejie Chi</a>, and <a href="https://www.andrew.cmu.edu/user/beidic/">Beidi Chen</a>
<br><br>
<p><i>ArXiv</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/pdf/2410.21465.pdf" target="blank">arXiv</a> / <a href="https://bytedance.github.io/ShadowKV/" target="blank">website</a> / <a href="https://github.com/bytedance/ShadowKV" target="blank">code <img src="https://img.shields.io/github/stars/bytedance/ShadowKV?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/sun2024shadowkv.txt", target="blank">bibtex</a>
</p>
<br>
<p>High-Throughput Long-Context LLM Inference System</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table>
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/specrej.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://Zanette-Labs.github.io/SpeculativeRejection" target="blank"><b>Fast Best-of-N Decoding via Speculative Rejection
</b></a><br>
<br>
<strong>Hanshi Sun*</strong>, Momin Haider*, <a href="https://rqzhangberkeley.github.io/">Ruiqi Zhang*</a>, Huitao Yang, <a href="https://ece.princeton.edu/people/jiahao-qiu">Jiahao Qiu</a>, <a href="https://mingyin0312.github.io/">Ming Yin</a>, <a href="https://mwang.princeton.edu/">Mengdi Wang</a>, <a href="https://people.eecs.berkeley.edu/~bartlett/">Peter Bartlett</a>, and <a href="https://azanette.com/">Andrea Zanette*</a> (* for core authors)
<br><br>
<p><i>Conference on Neural Information Processing Systems (NeurIPS)</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/pdf/2410.20290.pdf" target="blank">arXiv</a> / <a href="https://Zanette-Labs.github.io/SpeculativeRejection" target="blank">website</a> / <a href="https://github.com/Zanette-Labs/SpeculativeRejection" target="blank">code <img src="https://img.shields.io/github/stars/Zanette-Labs/SpeculativeRejection?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/sun2024fast.txt", target="blank">bibtex</a>
</p>
<br>
<p>Fast Inference-time Aligment Algorithm</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tr onmouseout="triforce_stop()" onmouseover="triforce_start()" bgcolor="#ffffd0">
<td style="padding:10px;width:25%;vertical-align:middle">
<div>
<div style="opacity: 0; position: absolute; transition: opacity .2s ease-in-out; -moz-transition: opacity .2 ease-in-out; -webkit-transition: opacity .2 ease-in-out;" id="triforce_image">
<video loop autoplay muted width="400px" >
<source src="assets/video/TriForce.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>
</div>
<div>
<img src="assets/img/TriForce.png" alt="TriForce", style="width: 400px; padding-top:40px; padding-bottom: 25px;">
</div>
</div>
<script type="text/javascript">
function triforce_start() {
document.getElementById('triforce_image').style.opacity = "1";
}
function triforce_stop() {
document.getElementById('triforce_image').style.opacity = "0";
}
triforce_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://infini-ai-lab.github.io/TriForce/" target="blank"><b><span class="TriForceFont" style="display: inline; font-size: 15pt;">*</span>TriForce: Lossless Acceleration of Long Sequence Generation with Hierarchical Speculative Decoding</b></a>
<br>
<br>
<strong>Hanshi Sun</strong>, <a href="https://dreaming-panda.github.io/">Zhuoming Chen</a>, <a href="https://xinyuyang.me/">Xinyu Yang</a>, <a href="https://yuandong-tian.com/">Yuandong Tian</a>, and <a href="https://www.andrew.cmu.edu/user/beidic/">Beidi Chen</a><br><br>
<p><i>Conference on Language Modeling (COLM)</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/abs/2404.11912.pdf" target="blank">arXiv</a> / <a href="https://infini-ai-lab.github.io/TriForce/" target="blank">website</a> / <a href="https://github.com/infini-ai-lab/TriForce/" target="blank">code <img src="https://img.shields.io/github/stars/infini-ai-lab/TriForce?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="https://youtu.be/vRAaAyjr6Jo" target="blank">demo</a> / <a href="bibtex/sun2024triforce.txt", target="blank">bibtex</a>
</p>
<br>
<p>Training-free Lossless Long Sequence Generation Acceleration</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table>
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/BMAD.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<b>BMAD: Benchmarks for Medical Anomaly Detection
</b> <br>
<br>
Jinan Bao, <strong>Hanshi Sun</strong>, Hanqiu Deng, Zhaoxiang Zhang, and <a href="https://www.ece.ualberta.ca/~xingyu/index.html">Xingyu Li</a>
<br><br>
<p><i>Computer Vision and Pattern Recognition (CVPR) Workshop</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/pdf/2306.11876.pdf" target="blank">arXiv</a> / <a href="https://github.com/DorisBao/BMAD" target="blank">code <img src="https://img.shields.io/github/stars/DorisBao/BMAD?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/bao2024bmad.txt", target="blank">bibtex</a>
</p>
<br>
<p>This benchmark encompasses six reorganized datasets from five medical domains (i.e. brain MRI, liver CT, retinal OCT, chest X-ray, and digital histopathology) and three key evaluation metrics, and includes a total of fourteen state-of-the-art AD algorithms.</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
</div>
<div id="full_publications" style="display: none;">
<table style="background-color: #ffffd0;">
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/shadowkv_sys.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://bytedance.github.io/ShadowKV/" target="blank"><b><img src="assets/img/ShadowKV.png" width="24px"> ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference</a>
</b> <br>
<br>
<strong>Hanshi Sun</strong>, <a href="https://lchang20.github.io/">Li-Wen Chang</a>, <a href="https://sites.google.com/view/wenleibao/">Wenlei Bao</a>, <a href="https://sizezheng.github.io/">Size Zheng</a>, <a href="https://zheng-ningxin.github.io/">Ningxin Zheng</a>, <a href="https://scholar.google.com/citations?user=ZMfk2F8AAAAJ&hl=zh-CN">Xin Liu</a>, <a href="https://www.andrew.cmu.edu/user/harryd/">Harry Dong</a>, <a href="https://users.ece.cmu.edu/~yuejiec/index.html">Yuejie Chi</a>, and <a href="https://www.andrew.cmu.edu/user/beidic/">Beidi Chen</a>
<br><br>
<p><i>ArXiv</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/pdf/2410.21465.pdf" target="blank">arXiv</a> / <a href="https://bytedance.github.io/ShadowKV/" target="blank">website</a> / <a href="https://github.com/bytedance/ShadowKV" target="blank">code <img src="https://img.shields.io/github/stars/bytedance/ShadowKV?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/sun2024shadowkv.txt", target="blank">bibtex</a>
</p>
<br>
<p>High-Throughput Long-Context LLM Inference System</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table>
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/specrej.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://Zanette-Labs.github.io/SpeculativeRejection" target="blank"><b>Fast Best-of-N Decoding via Speculative Rejection
</b></a><br>
<br>
<strong>Hanshi Sun*</strong>, Momin Haider*, <a href="https://rqzhangberkeley.github.io/">Ruiqi Zhang*</a>, Huitao Yang, <a href="https://ece.princeton.edu/people/jiahao-qiu">Jiahao Qiu</a>, <a href="https://mingyin0312.github.io/">Ming Yin</a>, <a href="https://mwang.princeton.edu/">Mengdi Wang</a>, <a href="https://people.eecs.berkeley.edu/~bartlett/">Peter Bartlett</a>, and <a href="https://azanette.com/">Andrea Zanette*</a> (* for core authors)
<br><br>
<p><i>Conference on Neural Information Processing Systems (NeurIPS)</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/pdf/2410.20290.pdf" target="blank">arXiv</a> / <a href="https://Zanette-Labs.github.io/SpeculativeRejection" target="blank">website</a> / <a href="https://github.com/Zanette-Labs/SpeculativeRejection" target="blank">code <img src="https://img.shields.io/github/stars/Zanette-Labs/SpeculativeRejection?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/sun2024fast.txt", target="blank">bibtex</a>
</p>
<br>
<p>Fast Inference-time Aligment Algorithm</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tr onmouseout="triforce_stop()" onmouseover="triforce_start()" bgcolor="#ffffd0">
<td style="padding:10px;width:25%;vertical-align:middle">
<div>
<div style="opacity: 0; position: absolute; transition: opacity .2s ease-in-out; -moz-transition: opacity .2 ease-in-out; -webkit-transition: opacity .2 ease-in-out;" id="triforce_image">
<video loop autoplay muted width="400px" >
<source src="assets/video/TriForce.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>
</div>
<div>
<img src="assets/img/TriForce.png" alt="TriForce", style="width: 400px; padding-top:40px; padding-bottom: 25px;">
</div>
</div>
<script type="text/javascript">
function triforce_start() {
document.getElementById('triforce_image').style.opacity = "1";
}
function triforce_stop() {
document.getElementById('triforce_image').style.opacity = "0";
}
triforce_stop()
</script>
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<a href="https://infini-ai-lab.github.io/TriForce/" target="blank"><b><span class="TriForceFont" style="display: inline; font-size: 15pt;">*</span>TriForce: Lossless Acceleration of Long Sequence Generation with Hierarchical Speculative Decoding</b></a>
<br>
<br>
<strong>Hanshi Sun</strong>, <a href="https://dreaming-panda.github.io/">Zhuoming Chen</a>, <a href="https://xinyuyang.me/">Xinyu Yang</a>, <a href="https://yuandong-tian.com/">Yuandong Tian</a>, and <a href="https://www.andrew.cmu.edu/user/beidic/">Beidi Chen</a><br><br>
<p><i>Conference on Language Modeling (COLM)</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/abs/2404.11912.pdf" target="blank">arXiv</a> / <a href="https://infini-ai-lab.github.io/TriForce/" target="blank">website</a> / <a href="https://github.com/infini-ai-lab/TriForce/" target="blank">code <img src="https://img.shields.io/github/stars/infini-ai-lab/TriForce?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="https://youtu.be/vRAaAyjr6Jo" target="blank">demo</a> / <a href="bibtex/sun2024triforce.txt", target="blank">bibtex</a>
</p>
<br>
<p>Training-free Lossless Long Sequence Generation Acceleration</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table>
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/BMAD.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<b>BMAD: Benchmarks for Medical Anomaly Detection
</b> <br>
<br>
Jinan Bao, <strong>Hanshi Sun</strong>, Hanqiu Deng, Zhaoxiang Zhang, and <a href="https://www.ece.ualberta.ca/~xingyu/index.html">Xingyu Li</a>
<br><br>
<p><i>Computer Vision and Pattern Recognition (CVPR) Workshop</i>, 2024</p><br>
<p>
<a href="https://arxiv.org/pdf/2306.11876.pdf" target="blank">arXiv</a> / <a href="https://github.com/DorisBao/BMAD" target="blank">code <img src="https://img.shields.io/github/stars/DorisBao/BMAD?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/bao2024bmad.txt", target="blank">bibtex</a>
</p>
<br>
<p>This benchmark encompasses six reorganized datasets from five medical domains (i.e. brain MRI, liver CT, retinal OCT, chest X-ray, and digital histopathology) and three key evaluation metrics, and includes a total of fourteen state-of-the-art AD algorithms.</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table>
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/combat_label_noise.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<b>Combating Medical Noisy Labels by Disentangled Distribution Learning and Consistency Regularization
</b> <br> <br>
<a href="https://scholar.google.co.uk/citations?user=EnDCJKMAAAAJ&hl=en">Yi Zhou</a>, <a href="https://huangleibuaa.github.io">Lei Huang</a>, <a href="https://taozh2017.github.io">Tao Zhou</a> and <strong>Hanshi Sun</strong><br><br>
<i>Future Generation Computer Systems (FGCS)</i>, 2023<br><br>
<p>
<a href="https://www.sciencedirect.com/science/article/pii/S0167739X22004228" target="blank">paper</a> / <a href="bibtex/zhou2023combating.txt", target="blank">bibtex</a>
</p>
<br>
<p>Disentangled distribution learning reduces effect of label uncertainty and ambiguity</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
<table>
<tr>
<td style="padding:10px;width:25%;vertical-align:middle">
<img src="assets/img/ecg.png" width="400px">
</td>
<td style="padding:20px;width:75%;vertical-align:middle">
<b>Arrhythmia Classifier Using Convolutional Neural Network with Adaptive Loss-aware Multi-bit Networks
Quantization
</b>
<br>
<strong>Hanshi Sun</strong>, Ao Wang, Ninghao Pu, Zhiqing Li, Junguang Huang, <a href="https://electronic.seu.edu.cn/lh/list.htm">Hao Liu</a> and Zhi Qi <br><br>
<i>ICAICE</i>, 2021<br><br>
<p>
<a href="https://preminstrel.github.io/ECG-Classification/" target="blank">website</a> /
<a href="https://ieeexplore.ieee.org/document/9797527" target="blank">paper</a> /
<a href="https://github.com/preminstrel/ECG-Classification" target="blank">code <img src="https://img.shields.io/github/stars/preminstrel/ECG-Classification?style=social&label=Github+Stars" style="vertical-align:middle"></a> / <a href="bibtex/sun2021arrhythmia.txt", target="blank">bibtex</a>
</p>
<br>
<p>Present a 1-D adaptive loss-aware quantization, achieving a high compression rate that reduces memory consumption by 23.36x</p>
</td>
</tr>
<tr> </tr>
<tr> </tr>
<tr> </tr>
</table>
</div>
<script>
function showFullPublications() {
var selectedPublicationsDiv = document.getElementById('selected_publications');
selectedPublicationsDiv.style.display = 'none';
var fullPublicationsDiv = document.getElementById('full_publications');
fullPublicationsDiv.style.display = 'block';
}
function showSelectedPublications() {
var fullPublicationsDiv = document.getElementById('full_publications');
fullPublicationsDiv.style.display = 'none';
var selectedPublicationsDiv = document.getElementById('selected_publications');
selectedPublicationsDiv.style.display = 'block';
}
function selectButton(buttonId) {
var button1 = document.getElementById('button1');
var button2 = document.getElementById('button2');
if (buttonId === 'button1') {
button1.classList.add('selected-button');
button2.classList.remove('selected-button');
} else if (buttonId === 'button2') {
button1.classList.remove('selected-button');
button2.classList.add('selected-button');
}
}
</script>
<h2>Education</h2>
<table>
<tbody>
<tr>
<td width="800">
<p><strong> Carnegie Mellon University, Pittsburgh, United States (Aug 2023 - Dec 2024)</strong></p>
<i>M.S. in Electrical & Computer Engineering</i>
<ul>
<li>Overall GPA: <font style="color: #bb0000;"> 4.0/4.0</font></li>
<li>Teaching Assistant @ Introduction to Deep Learning, Introduction to Machine Learning</li>
</ul>
</td>
<td width="120px">
<img src="assets/img/cmu_logo.jpeg" width="110px">
</td>
<td width="170px">
<img src="assets/img/cmu_logo_2.png" width="165px">
</td>
</tr>
<tr></tr>
</tbody>
</table>
<table>
<tbody>
<tr>
<td width="800">
<p><strong> Southeast University, Nanjing, China (Sep 2019 - Jul 2023)</strong></p>
<i>B.E. in Electronic Science and Technology</i>
<ul>
<li>Overall GPA: <font style="color: #bb0000;"> 3.96/4.0, 93.69/100 </font></li>
<li>2021 & 2022 China National Scholarship</li>
</ul>
</td>
<td>
<img src="assets/img/seu_l.jpeg" width="285px">
</td>
</tr>
<tr></tr>
</tbody>
</table>
<!-- <h2>Research Experience</h2>
<table>
<tbody>
<tr>
<td width="800">
<p><strong>University of Alberta, Edmonton, Canada (Jul 2022 - Oct 2022)</strong></p>
<i>Mitacs Globalink Research Internship</i>
<ul>
<li>Supervisor: Prof. <a href="https://www.ece.ualberta.ca/~xingyu/index.html">Xingyu Li</a>
</li>
<li>Project: Weak/self Supervised Learning for Anomaly Detection</li>
<li>
<a href="https://www.mitacs.ca/en/programs/globalink/globalink-graduate-fellowship">
Globalink Graduate Fellowship
</a>
</li>
</ul>
</td>
<td><br><br><br><br></td>
<td>
<img src="assets/img/University_of_Alberta_logo.png" width="285px">
<center><img src="assets/img/mitacs.jpeg" width="245px"></center>
</td>
</tr>
</tbody>
</table>
<table>
<tbody>
<tr>
<td width="800">
<p><strong> Southeast University, Nanjing, China (Oct 2020 - Jul 2022)</strong></p>
<i>Research Assitant in <a href="http://palm.seu.edu.cn/members.html">PALM Lab</a> and <a href="https://electronic.seu.edu.cn/dz_en/11384/list.htm">
National Engineering Research Center For ASIC
</a></i>
<ul>
<li>Supervisor: Prof.
<a href="https://scholar.google.co.uk/citations?user=EnDCJKMAAAAJ&hl=en">Yi Zhou</a>, Prof. <a href="https://electronic.seu.edu.cn/lh/list.htm">Hao Liu</a>
</li>
<li>Project 1: Energy-efficient DNN Algorithm for ECG Classification</li>
<li>Project 2: Multi-Modal Multi-task Transformer on Disease Detection</li>
<li>Project 3: Multi-task Learning For Multi-disease Classification</li>
</ul>
</td>
<td width="141px">
<img id="school_logo" src="assets/img/PALMLogo.png" width="120px">
</td>
<td width="144px">
<img src="assets/img/seu_ee.jpeg" width="125px">
</td>
</tr>
<tr></tr>
</tbody>
</table> -->
<h2>Professional Experience</h2>
<table>
<tbody>
<tr>
<td width="800">
<p><strong>ByteDance Inc., Seattle, United States (Jun 2024 - Present)</strong></p>
<i>Machine Learning System Research Intern in Seed-Foundation Team</i>
<ul>
<li>Introduced ShadowKV, a high-throughput long-context LLM inference system</li>
<li>ShadowKV can support 6x larger batch sizes and boost throughput by 3.04x while maintaining accuracy on downstream tasks</li>
</ul>
</li>
</td>
<td width="288px">
<img src="assets/img/ByteDance Logo English.png" width="288px" style="padding:10px;vertical-align:middle">
</td>
</tr>
<tr>
<td width="800">
<p><strong>Apple Inc., Shenzhen, China (Nov 2022 - Jun 2023)</strong></p>
<i>R&D Intern in iPad System EE</i>
<ul>
<li>Built a python automation test frame that can run on multiple units, collect and analyze logs</li>
<li>Issue reproduction, symptom capture, and hands-on debugging for coexistence testing</li>
<li>Created web pages with diverse visualization of the analyzed data using Flask</li>
</ul>
</li>
</td>
<td width="288px">
<img src="assets/img/apple.svg" width="288px">
</td>
</tr>
<tr></tr>
</tbody>
</table>
<div clss="footer">
<p>
<center>
<div align="center" style="width:18%">
<script type="text/javascript" id="clstr_globe"
src="//clustrmaps.com/globe.js?d=XU6JrQnKZEyIB6O3OwVxgskULQ1Qf9Uu5D17OS6SISE"></script>
</div>
</center>
</p>
<center>
<p style="color: darkgray;"> © Hanshi Sun 2024</p>
</center>
</div>
</body>
</html>