-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
424 lines (338 loc) · 19.4 KB
/
index.html
File metadata and controls
424 lines (338 loc) · 19.4 KB
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
<!DOCTYPE html>
<html lang="en">
<!-- Head -->
<head> <!-- Metadata, OpenGraph and Schema.org -->
<!-- Website verification -->
<meta name="google-site-verification" content="" /><meta name="msvalidate.01" content="" />
<!-- Standard metadata -->
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<title>LINs Lab</title>
<meta name="author" content="LINs Lab at Westlake University (Since 2022.11)" />
<meta name="description" content="" />
<meta name="keywords" content="westlake, deep learning, optimization, learning, inference, distributed deep learning, systems" />
<!-- Bootstrap & MDB -->
<link href="https://cdn.jsdelivr.net/npm/bootstrap@4.6.1/dist/css/bootstrap.min.css" rel="stylesheet" integrity="sha256-DF7Zhf293AJxJNTmh5zhoYYIMs2oXitRfBjY+9L//AY=" crossorigin="anonymous">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/mdbootstrap@4.20.0/css/mdb.min.css" integrity="sha256-jpjYvU3G3N6nrrBwXJoVEYI/0zw8htfFnhT9ljN3JJw=" crossorigin="anonymous" />
<!-- Fonts & Icons -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@fortawesome/fontawesome-free@5.15.4/css/all.min.css" integrity="sha256-mUZM63G8m73Mcidfrv5E+Y61y7a12O5mW4ezU3bxqW4=" crossorigin="anonymous">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/academicons@1.9.1/css/academicons.min.css" integrity="sha256-i1+4qU2G2860dGGIOJscdC30s9beBXjFfzjWLjBRsBg=" crossorigin="anonymous">
<link rel="stylesheet" type="text/css" href="https://fonts.googleapis.com/css?family=Roboto:300,400,500,700|Roboto+Slab:100,300,400,500,700|Material+Icons">
<!-- Code Syntax Highlighting -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jwarby/jekyll-pygments-themes@master/github.css" media="none" id="highlight_theme_light" />
<!-- Styles -->
<link rel="shortcut icon" href="data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 viewBox=%220 0 100 100%22><text y=%22.9em%22 font-size=%2290%22>⚛️</text></svg>">
<link rel="stylesheet" href="/assets/css/main.css">
<link rel="canonical" href="https://lins-lab.github.io/">
<!-- Dark Mode -->
</head>
<!-- Body -->
<body class="fixed-top-nav ">
<!-- Header -->
<header>
<!-- Nav Bar -->
<nav id="navbar" class="navbar navbar-light navbar-expand-sm fixed-top">
<div class="container">
<!-- Navbar Toggle -->
<button class="navbar-toggler collapsed ml-auto" type="button" data-toggle="collapse" data-target="#navbarNav" aria-controls="navbarNav" aria-expanded="false" aria-label="Toggle navigation">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar top-bar"></span>
<span class="icon-bar middle-bar"></span>
<span class="icon-bar bottom-bar"></span>
</button>
<div class="collapse navbar-collapse text-right" id="navbarNav">
<ul class="navbar-nav ml-auto flex-nowrap">
<!-- About -->
<li class="nav-item active">
<a class="nav-link" href="/">About<span class="sr-only">(current)</span></a>
</li>
<!-- Other pages -->
<li class="nav-item ">
<a class="nav-link" href="/group/">Group</a>
</li>
<li class="nav-item ">
<a class="nav-link" href="/publications/">Publications</a>
</li>
<li class="nav-item ">
<a class="nav-link" href="/teaching/">Teaching</a>
</li>
<li class="nav-item ">
<a class="nav-link" href="/openings/">Open Positions</a>
</li>
</ul>
</div>
</div>
</nav>
</header>
<!-- Content -->
<div class="container mt-5">
<!-- about.html -->
<div class="post">
<header class="post-header">
<h1 class="post-title">
<!-- LINs Lab --><figure>
<picture>
<source media="(max-width: 480px)" srcset="/assets/img/logo-480.webp"></source>
<source media="(max-width: 800px)" srcset="/assets/img/logo-800.webp"></source>
<source media="(max-width: 1400px)" srcset="/assets/img/logo-1400.webp"></source>
<!-- Fallback to the original file -->
<img class="img-fluid z-dept-1 rounded" src="/assets/img/logo.png">
</picture>
</figure>
</h1>
<p class="desc">
Welcome to the website of the <b><u>L</u></b>earning and <b><u>IN</u></b>ference <b><u>S</u></b>ystems (<b>LINs</b>) <b>Lab</b> at Westlake University!
</p>
</header>
<article>
<div class="profile float-right">
</div>
<div class="clearfix">
<p>Our research interests lie in the intersection of optimization and generalization for deep learning:</p>
<ul>
<li>
<em>leveraging</em> theoretical/empirical understanding (e.g., loss landscape, and training dynamics)</li>
<li>
<em>to design</em> efficient & robust methods (both learning and inference)</li>
<li>
<em>for</em> deep learning (centralized) and collaborative deep learning (distributed and/or decentralized),</li>
<li>
<em>under</em> imperfect environments (e.g., noisy, heterogeneous, and hardware-constrained).</li>
</ul>
<p>Lab activities:</p>
<ul>
<li>We are running a research seminar on <em><a href="https://dlo-seminar.github.io/" target="_blank" rel="noopener noreferrer">Deep Learning and Optimization</a></em>.</li>
</ul>
</div>
<!-- News -->
<div class="news">
<br>
<h2>News</h2>
<div class="table-responsive">
<table class="table table-sm table-borderless">
<tr>
<th scope="row">Apr 28, 2026</th>
<td>
Together with <a href="https://github.com/inclusionAI" target="_blank" rel="noopener noreferrer">Inclusion AI</a>, we are pleased to announce the release of <a href="https://arxiv.org/abs/2604.20796" target="_blank" rel="noopener noreferrer">LLaDA2.0-Uni</a>.
</td>
</tr>
<tr>
<th scope="row">Feb 21, 2026</th>
<td>
Several papers from our group were accepted at this year’s <a href="https://cvpr.thecvf.com/" target="_blank" rel="noopener noreferrer">CVPR 2026</a> conference.
<ul>
<li>Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training</li>
<li>Dual-Granularity Memory for Efficient Video Generation</li>
<li>Exploring Spatial Intelligence from a Generative Perspective</li>
<li>Eliciting Complex Spatial Reasoning in MLLMs through Wide-Baseline Matching</li>
</ul>
</td>
</tr>
<tr>
<th scope="row">Jan 26, 2026</th>
<td>
Several papers from our group were accepted at this year’s <a href="https://iclr.cc/" target="_blank" rel="noopener noreferrer">ICLR 2026</a> conference.
Congratulations to our PhD student <a href="https://gyxxyg.github.io/yongxinguo/" target="_blank" rel="noopener noreferrer">Peng SUN</a>, <a href="https://zhenglin-cheng.com/" target="_blank" rel="noopener noreferrer">Zhenglin CHENG</a>, and <a href="https://siyuan.lu/" target="_blank" rel="noopener noreferrer">Siyuan LU</a>, and our internship student <a href="https://scholar.google.com/citations?user=8qv_yagAAAAJ&hl=en" target="_blank" rel="noopener noreferrer">Fulin LIN</a>.
<ul>
<li><a href="https://openreview.net/forum?id=GnawtLKGkP" target="_blank" rel="noopener noreferrer">Any-step Generation via N-th Order Recursive Consistent Velocity Field Estimation</a></li>
<li><a href="https://arxiv.org/abs/2512.05150" target="_blank" rel="noopener noreferrer">TwinFlow: Realizing One-step Generation on Large Models with Self-adversarial Flows</a></li>
<li><a href="https://arxiv.org/abs/2510.26585" target="_blank" rel="noopener noreferrer">Stop Wasting Your Tokens: Towards Efficient Runtime Multi-Agent Systems</a></li>
<li><a href="https://arxiv.org/abs/2510.10197" target="_blank" rel="noopener noreferrer">Don’t Just Fine-tune the Agent, Tune the Environment</a></li>
</ul>
</td>
</tr>
<tr>
<th scope="row">Dec 9, 2025</th>
<td>
We are excited to release TwinFlow (<a href="https://arxiv.org/abs/2512.05150" target="_blank" rel="noopener noreferrer">arxiv</a> and <a href="https://github.com/inclusionAI/TwinFlow" target="_blank" rel="noopener noreferrer">code</a>), a simple, flexible, and memory-efficient framework for one-step generation on large-scale models. The project has already garnered 200+ GitHub stars in a few days!
</td>
</tr>
<tr>
<th scope="row">Sep 18, 2025</th>
<td>
Our <a href="https://arxiv.org/abs/2505.20510" target="_blank" rel="noopener noreferrer">CPathAgent</a> was accepted at this year’s <a href="https://nips.cc" target="_blank" rel="noopener noreferrer">NeurIPS 2025</a> conference. Congratulations to Yuxuan.
</td>
</tr>
</table>
</div>
</div>
<!-- Selected papers -->
<div class="publications">
<h2>Selected publications</h2>
<ol class="bibliography">
<li>
<!-- _layouts/bib.html -->
<div class="row">
<div class="col-sm-2 abbr"></div>
<!-- Entry bib key -->
<div id="sun2025ucgm" class="col-sm-8">
<!-- Title -->
<div class="title">Unified Continuous Generative Models</div>
<!-- Author -->
<div class="author">
<a href="https://sp12138.github.io/" target="_blank" rel="noopener noreferrer">Sun, Peng</a>, Jiang, Yi, and <a href="https://tlin-taolin.github.io/" target="_blank" rel="noopener noreferrer">Lin, Tao</a>
</div>
<!-- Journal/Book title and date -->
<div class="periodical">
<em>(arXiv:2505.07447)</em> preprint
</div>
<!-- Links/Buttons -->
<div class="links">
<a href="https://arxiv.org/abs/2505.07447" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">PDF</a>
<a href="https://github.com/LINs-lab/UCGM" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">Code</a>
</div>
</div>
</div>
</li>
<li>
<!-- _layouts/bib.html -->
<div class="row">
<div class="col-sm-2 abbr"><abbr class="badge">ICLR 2026</abbr></div>
<!-- Entry bib key -->
<div id="cheng2025twinflow" class="col-sm-8">
<!-- Title -->
<div class="title">TwinFlow: Realizing One-step Generation on Large Models with Self-adversarial Flows</div>
<!-- Author -->
<div class="author">
<a href="https://zhenglin-cheng.com/" target="_blank" rel="noopener noreferrer">Cheng, Zhenglin*</a>, <a href="https://sp12138.github.io/" target="_blank" rel="noopener noreferrer">Sun, Peng*</a>, <a href="https://scholar.google.com/citations?user=n44GlFcAAAAJ&hl=en" target="_blank" rel="noopener noreferrer">Li, Jianguo</a>, and <a href="https://tlin-taolin.github.io/" target="_blank" rel="noopener noreferrer">Lin, Tao</a>
</div>
<!-- Journal/Book title and date -->
<div class="periodical">
<em>In International Conference on Learning Representations (ICLR),</em> 2026
</div>
<!-- Links/Buttons -->
<div class="links">
<a href="https://arxiv.org/abs/2512.05150" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">PDF</a>
<a href="https://github.com/inclusionAI/TwinFlow" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">Code</a>
</div>
</div>
</div>
</li>
<li>
<!-- _layouts/bib.html -->
<div class="row">
<div class="col-sm-2 abbr"><abbr class="badge">ICLR 2025</abbr></div>
<!-- Entry bib key -->
<div id="yao2024deft" class="col-sm-8">
<!-- Title -->
<div class="title">DeFT: Decoding with Flash Tree-Attention for Efficient Tree-structured LLM Inference</div>
<!-- Author -->
<div class="author">
<a href="https://scholar.google.com/citations?user=idZOK_kAAAAJ&hl=en" target="_blank" rel="noopener noreferrer">Yao, Jinwei*</a>, Chen, Kaiqi*, <a href="https://zkx06111.github.io/" target="_blank" rel="noopener noreferrer">Zhang, Kexun</a>, <a href="https://cs.stanford.edu/~jiaxuan/" target="_blank" rel="noopener noreferrer">You, Jiaxuan</a>, <a href="https://binhangyuan.github.io/site/" target="_blank" rel="noopener noreferrer">Yuan, Binhang</a>, <a href="https://wangzeke.github.io/" target="_blank" rel="noopener noreferrer">Wang, Zeke</a>, and <a href="https://tlin-taolin.github.io/" target="_blank" rel="noopener noreferrer">Lin, Tao</a>
</div>
<!-- Journal/Book title and date -->
<div class="periodical">
<em>In International Conference on Learning Representations (ICLR), Spotlight, abridged in ICLR workshop AGI (Oral),</em> 2025
</div>
<!-- Links/Buttons -->
<div class="links">
<a href="https://arxiv.org/abs/2404.00242" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">PDF</a>
<a href="https://github.com/LINs-lab/DeFT" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">Code</a>
</div>
</div>
</div>
</li>
<li>
<!-- _layouts/bib.html -->
<div class="row">
<div class="col-sm-2 abbr"><abbr class="badge">NeurIPS 2024</abbr></div>
<!-- Entry bib key -->
<div id="sun2024efficiency" class="col-sm-8">
<!-- Title -->
<div class="title">Efficiency for Free: Ideal Data Are Transportable Representations</div>
<!-- Author -->
<div class="author">
<a href="https://sp12138.github.io/" target="_blank" rel="noopener noreferrer">Sun, Peng</a>, Jiang, Yi, and <a href="https://tlin-taolin.github.io/" target="_blank" rel="noopener noreferrer">Lin, Tao</a>
</div>
<!-- Journal/Book title and date -->
<div class="periodical">
<em>In Advances in Neural Information Processing Systems (NeurIPS),</em> 2024
</div>
<!-- Links/Buttons -->
<div class="links">
<a href="https://arxiv.org/abs/2405.14669" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">PDF</a>
<a href="https://github.com/LINs-lab/ReLA" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">Code</a>
</div>
</div>
</div>
</li>
<li>
<!-- _layouts/bib.html -->
<div class="row">
<div class="col-sm-2 abbr"><abbr class="badge">ICLR 2020</abbr></div>
<!-- Entry bib key -->
<div id="lin2020dont" class="col-sm-8">
<!-- Title -->
<div class="title">Don’t Use Large Mini-batches, Use Local SGD</div>
<!-- Author -->
<div class="author">
<a href="https://tlin-taolin.github.io/" target="_blank" rel="noopener noreferrer">Lin, Tao</a>, <a href="https://www.sstich.ch/" target="_blank" rel="noopener noreferrer">Stich, Sebastian U</a>, <a href="https://kishinmh.github.io/" target="_blank" rel="noopener noreferrer">Patel, Kumar Kshitij</a>, and <a href="https://scholar.google.com/citations?user=r1TJBr8AAAAJ&hl=en" target="_blank" rel="noopener noreferrer">Jaggi, Martin</a>
</div>
<!-- Journal/Book title and date -->
<div class="periodical">
<em>In International Conference on Learning Representations (ICLR),</em> 2020
</div>
<!-- Links/Buttons -->
<div class="links">
<a href="https://arxiv.org/abs/1808.07217" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">PDF</a>
<a href="https://github.com/epfml/LocalSGD-Code" class="btn btn-sm z-depth-0" role="button" target="_blank" rel="noopener noreferrer">Code</a>
</div>
</div>
</div>
</li>
</ol>
</div>
<!-- Social -->
<div class="social">
<div class="contact-icons">
<a href="mailto:%69%74%61%6D%74%61%6F@%67%6D%61%69%6C.%63%6F%6D" title="email"><i class="fas fa-envelope"></i></a>
<a href="https://github.com/lins-lab" title="GitHub" target="_blank" rel="noopener noreferrer"><i class="fab fa-github"></i></a>
</div>
<div class="contact-note">
</div>
</div>
</article>
</div>
</div>
<!-- Footer -->
<footer class="fixed-bottom">
<div class="container mt-0">
© Copyright 2026 LINs Lab at Westlake University (Since 2022.11). Last updated: May 01, 2026.
</div>
</footer>
<!-- JavaScripts -->
<!-- jQuery -->
<script src="https://cdn.jsdelivr.net/npm/jquery@3.6.0/dist/jquery.min.js" integrity="sha256-/xUj+3OJU5yExlq6GSYGSHk7tPXikynS7ogEvDej/m4=" crossorigin="anonymous"></script>
<!-- Bootsrap & MDB scripts -->
<script src="https://cdn.jsdelivr.net/npm/@popperjs/core@2.11.2/dist/umd/popper.min.js" integrity="sha256-l/1pMF/+J4TThfgARS6KwWrk/egwuVvhRzfLAMQ6Ds4=" crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/bootstrap@4.6.1/dist/js/bootstrap.min.js" integrity="sha256-SyTu6CwrfOhaznYZPoolVw2rxoY7lKYKQvqbtqN93HI=" crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/mdbootstrap@4.20.0/js/mdb.min.js" integrity="sha256-NdbiivsvWt7VYCt6hYNT3h/th9vSTL4EDWeGs5SN3DA=" crossorigin="anonymous"></script>
<!-- Masonry & imagesLoaded -->
<script defer src="https://cdn.jsdelivr.net/npm/masonry-layout@4.2.2/dist/masonry.pkgd.min.js" integrity="sha256-Nn1q/fx0H7SNLZMQ5Hw5JLaTRZp0yILA/FRexe19VdI=" crossorigin="anonymous"></script>
<script defer src="https://cdn.jsdelivr.net/npm/imagesloaded@4/imagesloaded.pkgd.min.js"></script>
<script defer src="/assets/js/masonry.js" type="text/javascript"></script>
<!-- Medium Zoom JS -->
<script src="https://cdn.jsdelivr.net/npm/medium-zoom@1.0.6/dist/medium-zoom.min.js" integrity="sha256-EdPgYcPk/IIrw7FYeuJQexva49pVRZNmt3LculEr7zM=" crossorigin="anonymous"></script>
<script src="/assets/js/zoom.js"></script><!-- Load Common JS -->
<script src="/assets/js/common.js"></script>
<!-- MathJax -->
<script type="text/javascript">
window.MathJax = {
tex: {
tags: 'ams'
}
};
</script>
<script defer type="text/javascript" id="MathJax-script" src="https://cdn.jsdelivr.net/npm/mathjax@3.2.0/es5/tex-mml-chtml.js"></script>
<script defer src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-X5CE01R8Z8"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() { dataLayer.push(arguments); }
gtag('js', new Date());
gtag('config', 'G-X5CE01R8Z8');
</script>
</body>
</html>