Yilun Xu
@xuyilun2
Followers
1,280
Following
280
Media
13
Statuses
65
CS PhD Candidate @MIT_CSAIL . Research intern @NVIDIA , view are my own
Cambridge MA
Joined September 2018
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#23point5FinalEP
• 724864 Tweets
LOVTITUDE FANMEETING
• 553035 Tweets
FB MADE YOU BLUSH
• 465475 Tweets
Xavi
• 246573 Tweets
Kabosu
• 183752 Tweets
Flick
• 139408 Tweets
#한성수_방시혁이랑_손잡고_한강가
• 135570 Tweets
The ICJ
• 101385 Tweets
クロリンデ
• 97033 Tweets
シグウィン
• 90650 Tweets
LEAVE SEVENTEEN ALONE
• 88447 Tweets
Laporta
• 81993 Tweets
Ten Hag
• 73374 Tweets
ノーヒットノーラン
• 71517 Tweets
Memorial Day
• 50104 Tweets
アルハイゼン
• 42801 Tweets
#ساعه_استجابه
• 42227 Tweets
McKenna
• 40308 Tweets
ショートアニメ
• 33178 Tweets
#ガルクラ
• 31454 Tweets
Congratulations LISA
• 30700 Tweets
العدل الدوليه
• 28167 Tweets
تشافي
• 26551 Tweets
Coutinho
• 23105 Tweets
INEOS
• 19859 Tweets
Morgan Spurlock
• 16017 Tweets
Super Size Me
• 15540 Tweets
Knives Out
• 14187 Tweets
Wake Up Dead Man
• 11979 Tweets
ジュラシックワールド
• 11700 Tweets
Benoit Blanc
• 10307 Tweets
I curate a list of interesting ICLR submissions on Diffusion/SDE/ODE-based generative models. (). Welcome to add more 🧐
3
85
434
Excited to share our
#NeurIPS2022
paper "Poisson Flow Generative Models". The PFGM ODE achieves
- SOTA performance in (continuous) normalizing flow family
- Faster sampling speed than SDEs in diffusion models
Paper:
Code: …
1/n
4
34
218
Officially passed my PhD thesis defense today! I'm deeply grateful to my collaborators and friends for their support throughout this journey. Huge thanks to my amazing thesis committee: Tommi Jaakkola (advisor),
@karsten_kreis
, and
@phillip_isola
! 🎓✨
18
13
227
In diffusion models, samplers are primarily ODE-centric, overlooking slower stochastic methods. However, we show that stochastic sampler can outperform previous samplers on Stable Diffusion, if we use stochasticity correctly!
check out Restart Sampling:
2
35
176
Restart is accepted by
#NeurIPS23
, and gets deployed in the popular webui . Let's combine the best of SDE (better quality) and ODE (faster sampling) samplers!
In diffusion models, samplers are primarily ODE-centric, overlooking slower stochastic methods. However, we show that stochastic sampler can outperform previous samplers on Stable Diffusion, if we use stochasticity correctly!
check out Restart Sampling:
2
35
176
1
11
71
Presenting "Restart Sampling for Improving Generative Processes" @
#NeurlPS2023
today!
Poster
#808
5pm-7pm
Come by to chat about the fast sampling for diffusion models!
Restart is accepted by
#NeurIPS23
, and gets deployed in the popular webui . Let's combine the best of SDE (better quality) and ODE (faster sampling) samplers!
1
11
71
2
12
63
non-IID sampling can promote diversity / mitigate memorization in diffusion models!
New paper!🤗
Do all your samples from Stable Diffusion or Dall-E look very similar to each other? It turns out IID sampling is to blame! We study this problem and propose Particle Guidance, a technique to obtain diverse samples that can be readily applied to your diffusion model!
4
87
440
0
6
53
Check out our ICLR spotlight paper on constructing the **orthogonal classifiers** that enable/outperform baselines in three tasks :
- Controlled style transfer
- Domain adaptation with label shifts
- Fairness
Paper:
Code:
1
2
41
Excited to share our latest paper which establishes a duality between generative models and physical processes 😃.
Generative models have been inspired by physics, but Eureka-type “inspirations” are mysterious. Is there a systematic way to convert physical processes to generative models? The answer is yes! This will largely augment design space of generative models.
2
55
231
1
2
33
@kohjingyu
This may not necessarily be a binary problem (diffusion versus auto-regressive). It is indeed possible to integrate the strengths of the two into a single model through a novel training pipeline. Stay tuned for our new model designed to achieve this 🙂
3
1
31
Quanta magazine just released an article featuring our recent series of works on generative AI 😬
Researchers are exploring whether “physics-inspired generative models” might offer more transparent and effective forms of artificial intelligence. Steve Nadis reports:
1
42
170
0
0
18
@JosephJacks_
@nearcyan
Thanks! Our recent experiments shows that we can further achieve 100x - 200x speedup with no loss on image quality, with some improvements on sampling methods. Stay tuned 😉
0
1
16
Happy to see the great application of V-information in the NLP domain😋
🎉🎉Super thrilled that our paper on Understanding Dataset Difficulty with V-usable information received an outstanding paper award at
#ICML2022
!! 🥳Looking forward to the broader applications of this framework. It was a total delight working with my
@allen_ai
intern,
@ethayarajh
16
29
349
1
0
15
Very detailed and educational blog of PFGM!
Stable Diffusion runs on physics-inspired Deep Learning.
Researchers from MIT (first authors
@ZimingLiu11
and
@xuyilun2
) have recently unveiled a new physics-inspired model that runs even faster!
This introduction has everything you need to know 👇
2
34
165
0
1
12
@LucaAmb
@NeurIPSConf
The phase transition corresponds to the far/middle/near field studied in
0
3
12
In practice, Restart beats SDE and ODE samplers in speed and quality on CIFAR-10 and ImageNet-64. Additionally, Restart achieves a better balance of text-image alignment, visual quality, and diversity on large-scale text-to-image Stable Diffusion!
Code:
1
0
9
Inspired by the electric field in physics, we interpret the data points as electrical charges on the z = 0 hyperplane in a space augmented with an additional dimension z The electric field lines transform data distribution into a uniform distribution on the large hemisphere.
3/n
1
3
7
ODE samplers are fast but plateau in performance while SDE samplers deliver better samples at the cost of increased sampling time. We attribute this difference to sampling errors: ODE involve smaller discretization errors while stochasticity in SDE contracts accumulated errors.
1
0
6
We learn the high dimensional version of the electric field, termed Poisson field, by neural networks. Specifically, we use a large batch to calculate the normalized field.
4/n
1
1
5
We also design a backward ODE for sampling. The backward ODE transforms samples from uniform hemisphere to the data distribution.
5/n
1
1
5
0
0
4
Based on these findings, we propose a novel sampling algorithm called Restart in order to better balance discretization errors and contraction. The sampling method alternates between adding substantial noise in additional forward steps and strictly following a backward ODE.
1
0
4
Our method achieves SOTA results on CIFAR10 dataset in flow family, faster sampling speed than SDEs in score-based and diffusion models, and more robustness. It can also scale to higher resolution dataset, e.g. LSUN bedroom 256x256.
n/n
0
1
4
1
0
4
In our updated version, we will show that the Restart sampling can also produce better samples in low-NFE regime (~20 NFE) on benchmarks and SD.
1
0
3
I really like this new paper on "Usable Information under Computational Constraints". It generalizes Shannon information to consider the ease of making predictions using a particular representation.
by
@baaadas
,
@shengjia_zhao
,
@StefanoErmon
et al.
4
31
252
0
0
2
1
0
1
@madebyollin
Thanks for pointing this out, unfortunately that's a limitation on current method and one has to construct separate reference batches for different conditions. But please try STF if you can maintain these batches! (naive DSM is STF with batch size=1)
0
0
1
@YunzhuLiYZ
@UofIllinois
@uofigrainger
@IllinoisCS
@StanfordSVL
@jiajunwu_cs
@drfeifei
Congrats Yunzhu!!
1
0
1