Andrei Bursuc
@abursuc
Followers
7,075
Following
1,102
Media
644
Statuses
8,484
Research scientist @valeoai | Teaching @Polytechnique @ENS_ULM | Alumni @upb1818 @Mines_Paris @Inria @ENS_ULM
Paris, France
Joined November 2008
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Luka
• 307444 Tweets
Kyrie
• 76798 Tweets
Gobert
• 69095 Tweets
Dallas
• 65274 Tweets
Mavs
• 60078 Tweets
छठे चरण
• 42126 Tweets
Anthony Edwards
• 25827 Tweets
Naz Reid
• 25414 Tweets
DPOY
• 24118 Tweets
サクナヒメ
• 21465 Tweets
#loksabhaelections24
• 21027 Tweets
Draymond
• 17865 Tweets
マイフリ
• 16272 Tweets
#超超超超ゲーマーズday1
• 15583 Tweets
Derbez
• 14562 Tweets
清宮卒業
• 14291 Tweets
レイちゃん
• 13572 Tweets
クマフェス
• 12334 Tweets
日産スタジアム
• 11802 Tweets
トレジャー
• 10321 Tweets
Pinned Tweet
Slides and most videos from
@ICCVConference
🎭The Many Faces of Reliability of Deep Learning for Real-World Deployment🌍 tutorial are now up!
Feel free to reach out w/ any questions and suggestions
#ICCV2023
-📽️ videos:
- 💻slides:
Join us this Tuesday for our
@ICCVConference
tutorial on "The Many Faces of Reliability of Deep Learning for Real-World Deployment" prepared by
@SharonYixuanLi
,
@puneetdokania
,
@tuan_hung_vu
, Dengxin Dai, Patrick Pérez and
@abursuc
#ICCV2023
3
16
61
0
29
120
A visual exploration of Gaussian Processes: beautiful interactive plots and a brief tutorial to make GPs more approachable
3
122
434
Decoder Denoising Pretraining for Semantic Segmentation: A fun and simple idea for pre-training the decoder for semantic segmentation 1/
4
56
317
Our work on learning to perform semantic segmentation without human supervision by driving around cities made it to
#eccv2022
More info coming soon.
2
40
260
Slides for most talks at Good Citizen at CVPR workshop are up Lots of useful advice and experience for writing and reviewing papers, how to do good research and evaluation, talks, how to organise your time
#CVPR2018
0
103
236
The brilliant Little Book of Deep Learning by
@francoisfleuret
is here! 🤩
Hoping now for an autograph session at a CV/ML venue soon.
4
20
238
New and not-so-new computer vision geeks on the block rejoice: the 2nd edition of Rick Szeliski's famous book on Computer Vision Algorithms and Applications is up and free to download as PDF
2
40
179
So far,
@georgiagkioxari
's talk on "Apples and Oranges: research in academia vs. industry" is the funniest and most thought-provoking talk at
#CVPR2023
2
24
159
New work spearheaded by S. Gidaris on self-supervised learning: OBoW - Online Bag-of-Visual-Words Generation for Unsupervised Representation Learning
Paper:
Code:
🧵👇
1/N
6
31
156
I'm a fan of papers proposing new "baselines". The methods are usually simpler and reach decent performance (few points below the usually more complex SoTA methods), while enabling a different view over the problem at hand.
5
14
153
CLIP-based research is moving so fast that even authors cannot keep pace with arXiv 🙃There are currently 3 MaskCLIP papers out there (all published in major conferences):
-
-
-
2
32
155
For those tuning from home,
@YejinChoinka
's excellent keynote at
#CVPR2023
unravels and discusses many of the findings from this work on the limits of Transformers to Compositionality
2
39
144
DINO learned excellent visual representations with impressive generalization.
Ever since, many researchers have tried to outmatch (or at least match) that via various self-supervised and/or masked image modelling strategies without succeeding
1
14
144
TRADI: Tracking deep neural network weight distributions -- work with G. Franchi We’re proposing a cheap method for getting ensembles of networks from a single network training 1/
3
32
132
Got a submission rejected from ICML, then NeurIPS. We've improved it further and sent it to ICLR. I'm happy that the paper is even better and clearer now, but, boy, this game can be annoying. Did I miss any news lately? :)
1
0
129
There were so many cool works on multi-camera bird's-eye-view perception recently.
If you want to catch-up or just starting in the area, this figure by
@AdamWHarley
does an effective summary of main approaches.
Capture from the awesome Simple-BEV work:
0
27
118
No accepted papers to announce, but I'm happy to see that as reviewer I contributed to the improvement and acceptance of a paper
#NeurIPS2021
2
3
121
Nice writeup by Chelsea Finn on recent meta-learning and few-shot learning techniques
2
39
119
Detectron is FAIR's new platform for state-of-the-art object detection algorithms in caffe2. It includes Mask R-CNN and a rich model zoo. All under Apache2 license.
0
56
113
Nice trick for initializing last layer of a CNN for fine-tuning or for adding an extra-class after training: add L2-norm layer and use CNN output from new class sample(s) as weights for new neuron. Results are good from first steps.
1
45
112
The outstanding
#CVPR2023
keynote talk by
@YejinChoinka
on "2050: An AI odyssey: dark matter of intelligence" is up
For those tuning from home,
@YejinChoinka
's excellent keynote at
#CVPR2023
unravels and discusses many of the findings from this work on the limits of Transformers to Compositionality
2
39
144
1
26
113
tempoGAN: nice use of adversarial training for super-resolution of temporally consistent fluid flows using a Volumetric-GAN
1
40
99
New dataset INTERACTION: trajectories of traffic participants in interactive scenarios (eg roundabout, intersection, merging of lanes on highways) from US, Germany and China. It's useful for intention and behavior prediction, interactions between drivers.
2
43
102
Slides and videos from our
#eccv2022
tutorial "Self-Supervision on Wheels: Advances in Self-Supervised Learning from Autonomous Driving Data" are now available:
2
28
105
"to arXiv or not (during review)" dilemma for less known labs: if you do it, you risk reviewers seeing it & hesitating due to lower fame, if you don't, you risk concurrent work from bigger lab on related idea or new SotA getting public & reducing your chances at resubmission 🤷♂️
4
8
104
From this view,
@ykilcher
is currently the most selective venue in CV+ML with 1-2 selected papers per week out of a pool of many 100s arXiv papers every week.
And he was saying that ICML acceptance rate was low 🙃
3
3
100
Since 1 week I'm no longer leaving the house without my DIY mask. I had initially bought in the "masks are not helpful" argument that lead to a lack of action from my side. Thanks to
@jeremyphoward
and
@math_rachel
for convincing me otherwise
#masks4all
2
14
98
The Grand Slam of the relentless computer vision researcher: submit a paper to CVPR -> ICCV/ECCV -> WACV/3DV hoping it will eventually get in that year.
3
6
103
After a time of relatively slow progress in core ResNet backbone, we're seeing now a mini Cambrian explosion of architectures: ViT, NFNet, RepVGG, LambdaNet, CaiT RedNet, BotNet, HaloNet, MLP-Mixer, etc.
Luckily tireless
@wightmanr
keeps track of all 🙏
3
16
100
Kudos to
#cvpr2019
Program Chairs for preparing a super-handy reviewer tutorial, including a summary of the decision process (which is quite opaque to many reviewers), some annotated good and bad reviews, and a few tips
0
43
97
Fun and insightful paper and post from Uber studying the Lottery Ticket Hypothesis Signs of the weights count the most for the init after pruning and one can find a masking of the weights with good performances using random weights - the Supermasks.
0
23
98
An excellent survey on image-based object localization w/o human supervision.
It's my go-to to catch up on the many new works in the area, here clearly organized and analyzed.
Well crafted, it can also serve as a template for next surveys.
Nice work by
@oriane_simeoni
& team👇
📢[Survey 📚] Object localization in images with zero manual annotation?🤩
➡️ We propose a survey discussing recent works exploiting ❄️self-supervised ViTs (incl. ICCV’23 & NeurIPS’23 works💫) w/
@EloiZablocki
@SpyrosGidaris
@gillespuy
& P. Pérez.
📄
2
17
91
1
17
96
In the past years we've seen some successful transfers of ideas from NLP to CV: Bag-of-Visual Words, skip-connections in LSTMs to ResNets, ViT from Transformers. Do you have some good examples of idea transfer from CV to NLP?
10
9
93
0/ Discover LOST, our unsupervised object discovery method based on self-supervised transformers. While simple, quick to install and very efficient, LOST improves SoTA by several points.
#bmvc2021
Paper:
Code:
3
21
91
New release of 's advanced DL course. Content is great w/ incredible amount of very recent stuff. I find myself recommending this course all the time to bootstrap people into DL and good coding practices for it. Congrats
@math_rachel
@jeremyphoward
Launching Cutting Edge Deep Learning for Coders: 2018 edition
I know a lot of folks have been waiting for this - hope it meets your expectations!
24
328
924
1
22
88
In spite of the drama around
#NIPS2018
sold-out, the organisers did well to release only 2.5k spots and reserve the others (up to 6-8k) to authors of accepted papers, top reviewers and workshop papers. There was a hick-up though with the delayed announcement of accepted papers
4
28
80
Implementing matrix decompositions (Cholesky, LQ, sym eigen) as differentiable operators
1
35
82
Using CNNs for explicit memorization of training data such that it can infer if an image was used for training, find which dataset was used for training, and if a validation image leaked during training Great work by
@alexsablay
@hjegou
et al.
0
25
86
CVPR deadline week
Stress level: 8/10
Home-alone father of 3 minions handling evening and morning routine, meals and taking kids to/back from school on time in different parts of town and then catch train to office
Stress level: ... Ahahahahaha! 🥲
5
1
82
The video tutorial on Capsule Networks from
@aureliengeron
is so good that even Hinton praises it
0
35
81
An open letter from Russian scientists and science journalists against the war with Ukraine (google translate capture 👇): "The responsibility for unleashing a new war in Europe lies entirely with Russia. There is no rational justification for this war."
0
12
76
Happy to announce relatively recent work with
@GianniFranchi10
on efficient BNNs with fewer assumptions on weights, amenable to complex CV architectures (DeepLV3+ w/ ResNet50) and tasks (semantic segmentation)
Paper:
Code:
🧵1/
2
19
79
The release of LLaMa-2 from
@MetaAI
with open weights and free for both research and commercial use, is definitely a seismic moment with ripples in academia, commercial applications, startups for LLMs or ML in general.
I was not expecting such a release this soon.
Meta releases Llama 2: Open Foundation and Fine-Tuned Chat Models
paper:
blog:
develop and release Llama 2, a collection of pretrained and fine-tuned large language models (LLMs) ranging in scale from 7 billion to 70 billion
35
575
2K
2
13
78
An inspiring talk for all advisors from Kilian Weinberger
@MLRetrospective
workshop
#neurips2020
Using 3 of his recent papers as example, he argues on the importance on focusing on the gained insight and not just settle for beating SoTA enough to get you a paper 1/
1
17
72
I've just received a copy of
@TacoCohen
's super thesis and I'm looking forward to dive in. Such manuscripts are really handy for kickstarting you in a field and I'm sure I'll recommend it to students and collaborators in the future. Kudos!
1
4
75
Battle of Backbones (BoB): Besides the awesome name, this paper is highly insighftul in comparing lots and lots of pretrained models on several various computer vision downstream tasks (my kind favorite way to analyze and understand models in addition to many metrics).
🚨Excited to announce a large-scale comparison of pretrained vision backbones including SSL, vision-language models, and CNNs vs ViTs across diverse downstream tasks ranging from classification to detection to OOD generalization and more! NeurIPS 2023🚨🧵
6
93
415
1
9
78
MAML(s) for the masses: a new pytorch library for implementing existing and developing new meta-learning algos The source papers features a pedagogical description of inner loop meta-learning algos
Happy to announce our paper on Generalized Inner Loop Meta Learning, aka Gimli (), with
@brandondamos
,
@denisyarats
, Phu Mon Htut, Artem Molchanov, Franziska Meier,
@douwekiela
,
@kchonyc
, and
@soumithchintala
. THREAD [1/6]
5
84
282
0
14
75
OpenOOD: Benchmarking Generalized Out-of-Distribution Detection by
@JingkangY
@SharonYixuanLi
@DanHendrycks
et al.
#NeurIPS2022
What a crazy effort implement and compare 35 ODD methods under 9 benchmarks!
paper:
code:
1
14
73
Student registration fees for ML conferences are a good bargain this year: ICML $25, COLT $30, AISTATS €40. While virtual conferences are less exciting than in-person ones,
they can put you on track to watch talks, read papers and talk with authors (uncrowded poster sessions)
2
13
71
We're organizing a workshop on Uncertainty Quantification for Computer Vision
@ICCVConference
with a fantastic line-up of speakers, talks from contributed papers and a competition on uncertainty quantification for autonomous driving
#iccv2023
Please spread the news.
0
18
72
Torch code for Deep Networks with Stochastic Depth is already out
0
36
71
The neural networks behind Google Voice transcription: from GMMs to LSTM RNNs
http://t.co/GJDcyD1kXK
0
37
66
This new augmented visualization system for fencing games is mesmerizing. Great work by
@kcimc
@rhizomatiks
and collaborators
3
13
65
Happy to see that Paris is becoming a hotspot for AI folks from academia, big tech or startups with a particular vision for openness and collaboration.
Kudos to
@ylecun
his relentless quest over the years in advertising Paris as a excellent place for that.
Open source AI is the way to go!
Proud to see
@huggingface
,
@scaleway
, &
@meta
joining to launch an AI startup accelerator at Station F.
This will help concretize our common vision of an open and collaborative AI ecosystem.
More from TechCrunch:
36
192
1K
0
6
68
Great blog-post by
@lilianweng
on recent progress in meta-learning (aka learning to learn fast, low-shot learning) covering the terminology and main families of approaches
0
23
66
Thrilled w/ CLIP-DINOiser by
@mkwysoczanska
et al. to compute dense CLIP features.
CLIP's limitation to global features-only annoyed many vision folks.
MaskCLIP (simple & fast) can get dense features, but noisy. We extend MaskCLIP w/ a few nudges from DINO localization priors 👇
🚨Happy to release on arXiv CLIP-DINOiser: Teaching CLIP a few DINO tricks🦖🎓
We obtain dense CLIP features in 1 forward pass w/o feature alteration and w/ almost no computational extra cost to facilitate open-vocabulary semantic segmentation 🧶
🖥️: [1/N]
4
37
217
1
15
66
This dataset does not exist: training models from generated images -- new work with V. Besnier and team Recent GANs, e.g. BigGAN, output increasingly high-quality images. What if we trained classifiers over fake images only, an initial aim of GANs? 1/
1
20
64
CelebA: a large-scale dataset of celebrity faces with attributes annotations - 10k celebs, 200k imgs, 40 attr/img
http://t.co/IhqX1X3hea
0
39
57
So
@ajmooch
's NFNets (Normalizing-Free Nets) are back in the spotlight w/ this mythbusting comparison of scaling properties against ViTs.
NFNets can scale well to JFT-4B level w/ predictable learning rate behavior.
Whether ConvNet or ViT, at scale, there's no BatchNorm.
ConvNets Match Vision Transformers at Scale
paper page:
Many researchers believe that ConvNets perform well on small or moderately sized datasets, but are not competitive with Vision Transformers when given access to datasets on the web-scale. We
20
178
881
2
17
63
I dare a challenge for
#cvpr2021
twitter: for each paper of yours that you advertise on twitter, please share 1-3 interesting papers from other teams. Authors will appreciate it and the community will be stronger. Here are some champions:
@ducha_aiki
@CSProfKGD
@artsiom_s
3
4
60
New GraphCut layer in PyTorch for differentiable submodular optimization proposed by Djolonga and Krause
#NIPS2017
1
21
60
Oh boy! Faster R-CNN and Mask R-CNN code for PyTorch 1.0 from Facebook folks. It's 2x faster than Detectron. Welcome to a new level of fun for future deep learning courses
1
10
59
Fast Language-Image Pre-training via Masking (FLIP)
It turns out that randomly masking out image patches with high mask ratio and encoding of visible patches only improves significantly the scalability and accuracy of CLIP (heavily used in so many tasks)
2
14
61
Two new posts from Vincent Vanhoucke on managing research teams. They're quite insightful for non-manager researchers as it covers typical time wasting practices and advice for guiding research projects
1
20
58
Fun paper towards more practical vision-based (multi-camera) perception for autonomous driving. Most benchmarks focus on accuracy often ignoring latency. Here, the authors evaluate online performance of SoTA methods with the inference delay impact:
4
6
60
Join us this Tuesday for our
@ICCVConference
tutorial on "The Many Faces of Reliability of Deep Learning for Real-World Deployment" prepared by
@SharonYixuanLi
,
@puneetdokania
,
@tuan_hung_vu
, Dengxin Dai, Patrick Pérez and
@abursuc
#ICCV2023
3
16
61
Join our
#cvpr2020
tutorial "Towards Annotation-Efficient Learning: Few-Shot, Self-Supervised, and Incremental Learning Approaches" with
@inthebrownbag
,
@relja_work
, S. Gidaris and me. We will be live on YouTube on June 19th 08:30 PDT / 17:30 CET
1
13
55
This is among my favorite papers at
#eccv2020
so far. A simple, elegant and effective approach. PyTorch code for the loss is up
Come join us as
#ECCV2020
for our Q&A sessions at 14:00 Tuesday and 00:00 Wednesday, UK Time!
We will discuss "Smooth-AP: Smoothing the Path Towards Large-Scale Image Retrieval" with myself,
@WeidiXie
,
@VickyKalogeiton
and Andrew Zisserman
#VGGatECCV2020
@Oxford_VGG
1
15
64
2
13
56
Matchbox: a new pytorch package from
@jekbradbury
for automatic batching, padding and masking
0
25
51
Currently working on a PhD thesis proposal and I find the whole process so rewarding: brainstorming for directions, measuring progress, state and trends in an area, taking time to reflect on interesting avenues to take in the next few years as well as their potential applications
2
1
52
Optimizing the latent space of generative networks, ie train a generator w/o the need to learn a discriminator
0
15
51
This might not be obvious from the West, but surely clear in East Europe: Ukrainians are currently fighting for our freedom too, not just theirs.
0
8
47