-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
420 lines (395 loc) · 19.4 KB
/
Copy pathindex.html
File metadata and controls
420 lines (395 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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Future Machine Learning & Systems Lab (FutureMLS)</title>
<meta name="description" content="The Future Machine Learning & Systems (FutureMLS) Lab, founded by Zhongzhu Zhou, researches efficient machine learning and systems: LLM inference, quantization, attention, and hardware-aware ML systems." />
<meta name="author" content="FutureMLS Lab" />
<!-- Open Graph -->
<meta property="og:title" content="Future Machine Learning & Systems Lab (FutureMLS)" />
<meta property="og:description" content="Efficient machine learning and systems: LLM inference, quantization, attention, and hardware-aware ML systems." />
<meta property="og:type" content="website" />
<meta property="og:image" content="assets/img/logo.png" />
<link rel="icon" type="image/png" href="assets/img/logo.png" />
<!-- Fonts: serif display + clean sans body for an academic feel -->
<link rel="preconnect" href="https://fonts.googleapis.com" />
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
<link href="https://fonts.googleapis.com/css2?family=Source+Serif+4:opsz,wght@8..60,400;8..60,500;8..60,600;8..60,700&family=Inter:wght@400;500;600&display=swap" rel="stylesheet" />
<link rel="stylesheet" href="assets/css/style.css" />
</head>
<body>
<!-- ===== Navigation ===== -->
<header class="nav" id="nav">
<div class="nav__inner container">
<a class="nav__brand" href="#home" aria-label="FutureMLS Lab home">
<img src="assets/img/logo.png" alt="" class="nav__logo" />
<span class="nav__brand-text">FutureMLS Lab</span>
</a>
<nav class="nav__links" id="navLinks" aria-label="Primary">
<a href="#about">About</a>
<a href="#founder">Zhongzhu Zhou</a>
<a href="#research">Research</a>
<a href="news.html">News</a>
<a href="#team">Team</a>
<a href="#projects">Projects</a>
<a href="#contact">Contact</a>
<a href="https://github.com/FutureMLS-Lab" class="nav__ext" target="_blank" rel="noopener">GitHub</a>
</nav>
<button class="nav__toggle" id="navToggle" aria-label="Toggle menu" aria-expanded="false">
<span></span><span></span><span></span>
</button>
</div>
</header>
<!-- ===== Hero (lab name on top) ===== -->
<section class="hero" id="home">
<div class="container hero__inner">
<p class="hero__eyebrow">Future Open Source Research</p>
<div class="hero__heading">
<img src="assets/img/logo.png" alt="FutureMLS Lab logo" class="hero__logo" />
<h1 class="hero__title">Future Machine Learning<br />& Systems Lab</h1>
</div>
<p class="hero__lead">
We study how to make machine learning more capable, efficient, and practical at
scale through four connected directions: algorithm and architecture innovation,
systems optimization, quantization, and training-driven modeling.
</p>
<div class="hero__actions">
<a href="#research" class="btn btn--primary">Research</a>
<a href="#team" class="btn btn--ghost">Team</a>
<a href="https://github.com/FutureMLS-Lab" class="btn btn--ghost" target="_blank" rel="noopener">GitHub ↗</a>
</div>
<ul class="hero__topics" aria-label="Research topics">
<li>Efficient ML Algorithm</li>
<li>Efficient ML System</li>
<li>Quantization</li>
<li>Modeling</li>
</ul>
</div>
</section>
<!-- ===== About (Who we are) ===== -->
<section class="section" id="about">
<div class="container">
<div class="section__head">
<span class="eyebrow">About</span>
<h2 class="section__title">Who we are</h2>
</div>
<div class="prose">
<p class="prose__lead">
Welcome to the <strong>Future Machine Learning & Systems (FutureMLS) Lab</strong>.
We work at the intersection of machine learning and computer systems. Today's
foundation models are remarkably capable but costly to train and serve. Our mission
is to close the gap between rapidly growing model capability and the real-world cost
of deploying these models.
</p>
<p>
We pursue <em>algorithm-system co-design</em> across four connected themes:
Efficient ML Algorithm for algorithm and architecture innovation, Efficient ML
System for systems optimization, Quantization as a core research focus, and Modeling
for improving models through training. Our work spans the AI stack, from methods and
model design to kernels, runtimes, and serving systems, and is open-source,
reproducible, and built to be used.
</p>
<p class="prose__meta">
<span><strong>Founder & PI:</strong> Zhongzhu Zhou</span>
</p>
</div>
</div>
</section>
<!-- ===== About Zhongzhu Zhou (Founder) ===== -->
<section class="section section--alt" id="founder">
<div class="container">
<div class="section__head">
<span class="eyebrow">Founder</span>
<h2 class="section__title">About Zhongzhu Zhou</h2>
</div>
<div class="founder">
<aside class="founder__media">
<img class="founder__photo" src="assets/img/people/zhongzhu-zhou.png" alt="Zhongzhu Zhou" />
<p class="founder__name">Zhongzhu Zhou</p>
<p class="founder__alias">Charlie Zhou · /ZHONG-JOO JOH/</p>
<p class="founder__role">Founder & Principal Investigator</p>
<div class="founder__links">
<a href="https://zhongzhuzhou.org/" target="_blank" rel="noopener">Homepage ↗</a>
<a href="https://scholar.google.com/" target="_blank" rel="noopener">Google Scholar ↗</a>
<a href="https://github.com/FutureMLS-Lab" target="_blank" rel="noopener">GitHub ↗</a>
<a href="mailto:zhongzhu.zhou@sydney.edu.au">Email</a>
</div>
</aside>
<div class="founder__body prose">
<p class="prose__lead">
Zhongzhu Zhou (Charlie Zhou) is the founder and principal investigator of the
Future Machine Learning & Systems Lab. He is a Senior Research Scientist on the
Turbo Team at Together AI, and earned his Ph.D. at the School of Computer Science,
University of Sydney.
</p>
<p>
His research spans <strong>efficient machine learning and systems</strong> — from
pretraining quality to efficient algorithms and algorithm–system co-design that
bridges emerging ML/LLM methods and real-world applications, improving both
productivity (usable, robust stacks) and performance (throughput, memory, and
cost-efficiency). He received his B.Eng. (Hons) from Sun Yat-sen University, and has
interned at Dolby, the DeepSpeed team at Microsoft, and Tencent.
</p>
<p>
He leads projects across the lab's four themes, including
<a href="projects/oscar.html">OSCAR</a> (2-bit KV-cache quantization)
and <a href="projects/care.html">CARE</a> (covariance-aware Multi-Head Latent Attention).
</p>
</div>
</div>
</div>
</section>
<!-- ===== Research ===== -->
<section class="section" id="research">
<div class="container">
<div class="section__head">
<span class="eyebrow">Research</span>
<h2 class="section__title">What we work on</h2>
<p class="section__sub">Four directions, one goal: efficient and capable AI at scale.</p>
</div>
<div class="topics">
<article class="topic reveal">
<span class="topic__no">01</span>
<h3 class="topic__title">Efficient ML Algorithm</h3>
<p>Algorithm and architecture innovations that improve capability while reducing compute, memory, and deployment cost.</p>
</article>
<article class="topic reveal">
<span class="topic__no">02</span>
<h3 class="topic__title">Efficient ML System</h3>
<p>System-level optimizations that make efficient methods practical end-to-end, from kernels and runtimes to high-throughput serving.</p>
</article>
<article class="topic reveal">
<span class="topic__no">03</span>
<h3 class="topic__title">Quantization</h3>
<p>A core research focus on low-bit weight, activation, and KV-cache quantization that preserves accuracy while cutting memory and compute.</p>
</article>
<article class="topic reveal">
<span class="topic__no">04</span>
<h3 class="topic__title">Modeling</h3>
<p>Model improvement through training optimization, architecture design, and adaptation methods that make models stronger and easier to use.</p>
</article>
</div>
</div>
</section>
<!-- ===== News ===== -->
<section class="section section--alt" id="news">
<div class="container">
<div class="section__head">
<span class="eyebrow">News</span>
<h2 class="section__title">Recent updates</h2>
</div>
<ul class="news">
<li class="news__item reveal">
<span class="news__date">May 26, 2026</span>
<p class="news__text"><strong><a href="projects/oscar.html">OSCAR</a></strong> crosses 300 ★ on GitHub in its first week — thanks to the open-source community.</p>
</li>
<li class="news__item reveal">
<span class="news__date">May 19, 2026</span>
<p class="news__text"><strong><a href="projects/oscar.html">OSCAR</a></strong> released — 2-bit KV-cache serving at 2.28 effective bits/element with near-BF16 accuracy on Qwen3 and GLM-4.7.</p>
</li>
<li class="news__item reveal">
<span class="news__date">Apr 2026</span>
<p class="news__text"><strong><a href="projects/care.html">CARE</a></strong> presented at <strong>ICLR 2026</strong>: covariance-aware, rank-enhanced decomposition for Multi-Head Latent Attention.</p>
</li>
</ul>
<div class="news__more">
<a class="btn btn--ghost" href="news.html">View all news →</a>
</div>
</div>
</section>
<!-- ===== Team ===== -->
<section class="section" id="team">
<div class="container">
<div class="section__head">
<span class="eyebrow">Team</span>
<h2 class="section__title">People</h2>
<p class="section__sub">A small group of researchers and advisors building in the open.</p>
</div>
<h3 class="team__group">Advisors</h3>
<div class="people people--advisors">
<article class="person reveal">
<img class="person__photo" src="assets/img/people/xiaoxia-wu.jpg" alt="Xiaoxia Wu" />
<div class="person__info">
<h4 class="person__name">Xiaoxia Wu</h4>
<p class="person__role">Advisor</p>
<p class="person__bio">Principal Research scientist in efficient ML and low-bit quantization, with extensive work across the DeepSpeed and Together AI.</p>
<ul class="person__tags">
<li>Efficient ML Algorithm</li>
<li>Efficient ML System</li>
<li>Quantization</li>
</ul>
<div class="person__links">
<a href="https://scholar.google.com/citations?user=Ry0Bdt8AAAAJ" target="_blank" rel="noopener">Scholar</a>
<a href="https://xwushirley.github.io/" target="_blank" rel="noopener">Homepage</a>
</div>
</div>
</article>
<article class="person reveal">
<img class="person__photo" src="assets/img/people/shuaiwen-song.png" alt="Shuaiwen Leon Song" />
<div class="person__info">
<h4 class="person__name">Shuaiwen Leon Song</h4>
<p class="person__role">Advisor</p>
<p class="person__bio">Professor at the University of Sydney; high-performance computing and ML systems.</p>
<ul class="person__tags">
<li>Efficient ML System</li>
</ul>
<div class="person__links">
<a href="https://scholar.google.com/citations?user=vt_QcOMAAAAJ" target="_blank" rel="noopener">Scholar</a>
<a href="https://shuaiwen-leon-song.github.io/" target="_blank" rel="noopener">Homepage</a>
</div>
</div>
</article>
</div>
<h3 class="team__group">Members</h3>
<div class="people">
<article class="person reveal">
<img class="person__photo" src="assets/img/people/fengxiang-bie.png" alt="Fengxiang “Bobbie” Bie" />
<div class="person__info">
<h4 class="person__name">Fengxiang “Bobbie” Bie</h4>
<p class="person__role">Student Researcher</p>
<p class="person__bio">Efficient ML algorithms and speculative decoding; contributor to CARE.</p>
<ul class="person__tags">
<li>Efficient ML Algorithm</li>
<li>Speculator</li>
</ul>
<div class="person__links">
<a href="https://github.com/FutureMLS-Lab" target="_blank" rel="noopener">GitHub</a>
<a href="#" target="_blank" rel="noopener">Scholar</a>
</div>
</div>
</article>
<article class="person reveal">
<img class="person__photo" src="assets/img/people/ziyan-chen.png" alt="Ziyan Chen" />
<div class="person__info">
<h4 class="person__name">Ziyan Chen</h4>
<p class="person__role">Student Researcher</p>
<p class="person__bio">Efficient ML algorithms and KV-cache compression; contributor to OSCAR and CARE.</p>
<ul class="person__tags">
<li>Efficient ML Algorithm</li>
</ul>
<div class="person__links">
<a href="https://github.com/FutureMLS-Lab" target="_blank" rel="noopener">GitHub</a>
<a href="#" target="_blank" rel="noopener">Scholar</a>
</div>
</div>
</article>
<article class="person reveal">
<img class="person__photo" src="assets/img/people/ryan-wang.png" alt="Ryan Wang" />
<div class="person__info">
<h4 class="person__name">Ryan Wang</h4>
<p class="person__role">Student Researcher</p>
<p class="person__bio">Efficient ML algorithms and speculative decoding for large-scale machine learning.</p>
<ul class="person__tags">
<li>Efficient ML Algorithm</li>
<li>Speculator</li>
</ul>
<div class="person__links">
<a href="https://github.com/FutureMLS-Lab" target="_blank" rel="noopener">GitHub</a>
<a href="#" target="_blank" rel="noopener">Scholar</a>
</div>
</div>
</article>
<article class="person reveal">
<img class="person__photo" src="assets/img/people/zhizhou-sha.jpg" alt="Zhizhou Sha" />
<div class="person__info">
<h4 class="person__name">Zhizhou Sha</h4>
<p class="person__role">Student Researcher</p>
<p class="person__bio">Quantization-aware training for efficient and accurate large language models.</p>
<ul class="person__tags">
<li>Quantization</li>
<li>Quantization-Aware Training</li>
</ul>
<div class="person__links">
<a href="https://github.com/JamesSand" target="_blank" rel="noopener">GitHub</a>
<a href="https://scholar.google.com/citations?user=eAObdYgAAAAJ" target="_blank" rel="noopener">Scholar</a>
<a href="https://jamessand.github.io/" target="_blank" rel="noopener">Homepage</a>
</div>
</div>
</article>
</div>
</div>
</section>
<!-- ===== Projects ===== -->
<section class="section section--alt" id="projects">
<div class="container">
<div class="section__head">
<span class="eyebrow">Projects</span>
<h2 class="section__title">Open-source research</h2>
<p class="section__sub">Selected projects from the lab — open the preview page for details, papers, and code.</p>
</div>
<div class="projects">
<a class="project reveal" href="projects/oscar.html">
<div class="project__preview">
<img src="assets/img/projects/oscar.png" alt="OSCAR preview" loading="lazy" />
</div>
<div class="project__body">
<div class="project__head">
<h3 class="project__name">OSCAR</h3>
<span class="project__tag">Quantization</span>
</div>
<p class="project__venue">2-bit KV-cache quantization · 2026</p>
<p class="project__desc">Attention-aware offline rotations compress the KV cache to 2.28 bits/element — ~8× memory reduction and up to ~7× higher throughput with near-BF16 accuracy.</p>
<span class="project__link">View project →</span>
</div>
</a>
<a class="project reveal" href="projects/care.html">
<div class="project__preview">
<img src="assets/img/projects/care.jpg" alt="CARE preview" loading="lazy" />
</div>
<div class="project__body">
<div class="project__head">
<h3 class="project__name">CARE</h3>
<span class="project__tag">Efficient ML Algorithm</span>
</div>
<p class="project__venue">Multi-Head Latent Attention · ICLR 2026</p>
<p class="project__desc">Covariance-aware low-rank decomposition that converts pretrained GQA/MHA into MLA — up to 215× lower one-shot perplexity at matched KV budgets.</p>
<span class="project__link">View project →</span>
</div>
</a>
</div>
</div>
</section>
<!-- ===== Contact ===== -->
<section class="section" id="contact">
<div class="container">
<div class="section__head">
<span class="eyebrow">Contact</span>
<h2 class="section__title">Get in touch</h2>
</div>
<div class="prose">
<p class="prose__lead">
We welcome collaborators, prospective students, and contributors who care about
efficient, open machine learning and systems.
</p>
<p class="contact__actions">
<a href="mailto:zhongzhu@futuremls.org" class="btn btn--primary">Email the lab</a>
<a href="https://github.com/FutureMLS-Lab" class="btn btn--ghost" target="_blank" rel="noopener">View our GitHub ↗</a>
</p>
</div>
</div>
</section>
<!-- ===== Footer ===== -->
<footer class="footer">
<div class="container footer__inner">
<div class="footer__brand">
<img src="assets/img/logo.png" alt="" class="footer__logo" />
<div>
<strong>FutureMLS Lab</strong>
<p>Future Machine Learning & Systems Lab</p>
</div>
</div>
<nav class="footer__links" aria-label="Footer">
<a href="#about">About</a>
<a href="#founder">Zhongzhu Zhou</a>
<a href="#research">Research</a>
<a href="#team">Team</a>
<a href="#projects">Projects</a>
<a href="https://github.com/FutureMLS-Lab" target="_blank" rel="noopener">GitHub</a>
</nav>
<p class="footer__copy">© <span id="year">2026</span> Future Machine Learning & Systems Lab. Founded by Zhongzhu Zhou.</p>
</div>
</footer>
<script src="assets/js/main.js"></script>
</body>
</html>