-
Notifications
You must be signed in to change notification settings - Fork 0
/
sampler.html
245 lines (224 loc) · 9.71 KB
/
sampler.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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="Training samplers with off policy RL.">
<meta name="keywords" content="diffusion models, gflownets, fine-tuning">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Off Policy Diffusion Samplers</title>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/images/favicon.svg">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/fontawesome.all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
<script type="text/x-mathjax-config">
MathJax.Hub.Config({tex2jax: {inlineMath: [['$','$'], ['\\(','\\)']]}});
</script>
<script type="text/javascript"
src="http://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS-MML_HTMLorMML">
</script>
</head>
<body>
<nav class="navbar" role="navigation" aria-label="main navigation" style="background-color: #333; color: #fff;">
<div class="navbar-brand">
<a role="button" class="navbar-burger" aria-label="menu" aria-expanded="false" style="color: #fff;">
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
</a>
</div>
<div class="navbar-menu">
<div class="navbar-start" style="flex-grow: 1; justify-content: center;">
<a class="navbar-item" href="#abstract" style="color: #fff; border-bottom: 0px solid #fff;">
Abstract
</a>
<a class="navbar-item" href="#setup" style="color: #fff; border-bottom: 0px solid #fff;">
Setup
</a>
<a class="navbar-item" href="#method" style="color: #fff; border-bottom: 0px solid #fff;">
Method
</a>
<a class="navbar-item" href="#results" style="color: #fff; border-bottom: 0px solid #fff;">
Empirical Results
</a>
<a class="navbar-item" href="#BibTeX" style="color: #fff; border-bottom: 0px solid #fff;">
BibTeX
</a>
</div>
</nav>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">Improved off-policy training of diffusion samplers</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://x.com/MarcinSendera">Marcin Sendera</a><sup>1,2</sup>
</span>
<span class="author-block">
<a href="https://minsuukim.github.io">Minsu Kim</a><sup>1,3</sup>
</span>
<span class="author-block">
<a href="https://sarthmit.github.io">Sarthak Mittal</a><sup>1</sup>
</span>
<span class="author-block">
<a href="https://pablo-lemos.github.io">Pablo Lemos</a><sup>1,4,5,6</sup>
</span>
<br/>
<span class="author-block">
<a href="https://lucascimeca.com">Luca Scimeca</a><sup>1</sup>
</span>
<span class="author-block">
<a href="https://jarridrb.github.io">Jarrid Rector-Brooks</a><sup>1,6</sup>
</span>
<span class="author-block">
<a href="https://mila.quebec/en/directory/alexandre-adam">Alexandre Adam</a><sup>1,4</sup>
</span>
<span class="author-block">
<a href="https://yoshuabengio.org">Yoshua Bengio</a><sup>1,7</sup>
</span>
<span class="author-block">
<a href="https://malkin1729.github.io">Nikolay Malkin</a><sup>1</sup>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">Mila - Quebec AI Institute</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>Université de Montréal</span>
<span class="author-block"><sup>2</sup>Jagiellonian University</span>
<span class="author-block"><sup>3</sup>KAIST</span>
<br/>
<span class="author-block"><sup>4</sup>Ciela Institute</span>
<span class="author-block"><sup>5</sup>Center for Computational Astrophysics, Flatiron Institute</span>
<br/>
<span class="author-block"><sup>6</sup>Dreamfold</span>
<span class="author-block"><sup>7</sup>CIFAR AI Chair</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">
<b>NeurIPS 2024</b>
</span>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2402.05098"
class="external-link button is-normal is-rounded">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span>
<span class="link-block">
<a href="https://arxiv.org/abs/2402.05098"
class="external-link button is-normal is-rounded">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
<!-- Code Link. -->
<span class="link-block">
<a href="https://github.com/GFNOrg/gfn-diffusion"
class="external-link button is-normal is-rounded">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body">
<img src="./static/images/final_teaser_sampler_manywell.png"
alt="Teaser image."
class="teaser-image"/>
<!-- <h2 class="subtitle has-text-centered">-->
<!-- Put a new figure here -->
<!-- </h2>-->
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3" id="abstract">Abstract</h2>
<div class="content has-text-justified">
<p>
We study the problem of training diffusion models to sample from a distribution with a given unnormalized
density or energy function. We benchmark several diffusion-structured inference methods, including
simulation-based variational approaches and off-policy methods (continuous generative flow networks).
Our results shed light on the relative advantages of existing algorithms while bringing into question some
claims from past work. We also propose a novel exploration strategy for off-policy methods, based on local
search in the target space with the use of a replay buffer, and show that it improves the quality of
samples on a variety of target distributions. Our code for the sampling methods and benchmarks studied
is made public at <a href="https://github.com/GFNOrg/gfn-diffusion">
<span>(link)</span>
</a> as a base for future work on diffusion models for amortized inference.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
<hr>
<hr>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>
@inproceedings{
sendera2024improved,
title={Improved off-policy training of diffusion samplers},
author={Marcin Sendera and Minsu Kim and Sarthak Mittal and Pablo Lemos and Luca Scimeca and Jarrid Rector-Brooks and Alexandre Adam and Yoshua Bengio and Nikolay Malkin},
booktitle={The Thirty-eighth Annual Conference on Neural Information Processing Systems},
year={2024},
url={https://openreview.net/forum?id=vieIamY2Gi}
}
</code></pre>
<!-- journal = {ICML}, -->
</div>
</section>
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link"
href="https://arxiv.org/pdf/2405.20971.pdf">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://github.com/GFNOrg/diffusion-finetuning" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
<div class="columns is-centered">
<p>
<!-- Corresponding Authors: <a href="mailto:[email protected]">Anikait Singh</a>, <a href="mailto:[email protected]">Fahim Tajwar</a>.<br> -->
We thank the Nerfies Team for their <a href="https://github.com/nerfies/nerfies.github.io">website template</a>.
</p>
</div>
</div>
</footer>
</body>
</html>