A gunman has murdered ten people and injured over a dozen others in Monterey Park, where tens of thousands of Asian American families had gathered to celebrate the Lunar New Year that day. He is still at large.
This should be getting way more coverage in media. A terrible act.
Yifan Jiang’s Tweets
There’s no paper deadline on Christmas Eve. But yesterday was both the #ACL2023 deadline and the Chinese New Year’s Eve, the biggest get-together-with-family event for many ethnically Chinese people. I know organizers are doing hard work, but maybe more future D&I can be great🥰
2
21
292
Tim Cook shares video to mark Chinese New Year on social media appleinsider.com/articles/23/01 #Apple
1
3
6
We're trying something new at this year. In response to the ongoing AI art explosion, we're broadening CVPR's Demos session to include an "Art Gallery" for AI-powered or generative media (diffusion, NeRF, AR, anything). Info and instructions here: cvpr2023.thecvf.com/Conferences/20
1
59
280
Show this thread
yeah…IMO “strong reject” is only for those should be desk reject or with ethic problem. But I do found people abuse this just because they dont like the submission
That is disrespectful to authors’ hard effort
Quote Tweet
Once upon a time, "Strong reject" was only given to really broken papers with no easy fixes. These days, some reviewers use it very lightly in ML conferences. If you are one of them, please, reconsider. This hurts your reputation... ACs will think you are not a reasonable person!
1
Also wanted to try #chatgpt - A rap battle between photogrammetry and NeRFs.
Quite fancy that it could even deal with the acronym.
2
15
75
Show this thread
ChatGPT >> Grammarly.
14
35
492
Ok，another one. I asked chatgpt to help me write a cold email asking 2023 summer internship at
Quote Tweet
I asked @OpenAI chatgpt to write a paper about neural rendering
1
I'll add a #a4 paper to every of my posters. hopes more can join. Part of our community may be arrested by Chinese government. I need to admitq i'm not brave enough to do this in mainland china. at least I can show my attitude here.
#A4Revolution
Quote Tweet
Why you should care about ongoing 
protests?For years, has been using AI to censor msgs &posts, and prosecute ppl w/ diff political views,or ppl who just wanna end zero-covid. You should care b/c the cool things you created—transformers, CNNs—are likely held in evil’s hands. twitter.com/mengyer/status…
protests?For years, has been using AI to censor msgs &posts, and prosecute ppl w/ diff political views,or ppl who just wanna end zero-covid. You should care b/c the cool things you created—transformers, CNNs—are likely held in evil’s hands. twitter.com/mengyer/status…1
10
79
Show this thread
1. What would you do if an #iclr2023 reviewer ignored the science of your "Markov Chain Monte Carlo" paper, but would give you the lowest score (1) solely because "chains" may refer to slavery, and because linking slavery to the fine people of Monte Carlo is offensive to them?
41
62
585
Show this thread
NeRF-SOS: Any-View Self-supervised Object Segmentation on Complex Scenes
Authors: Zhiwen Fan, Peihao Wang, Yifan Jiang, Xinyu Gong, Dejia Xu, Zhangyang Wang
arxiv.org/abs/2209.08776
#neuralfieldsoftheday
12
33
We have some research intern positions opening for summer 2023 in Adobe, working on exciting image editing/compositing including image harmonization, object relighting, object shadow generation and inverse rendering for image editing. Please feel free to contact me !
1
3
8
Imagine if Columbia reTweeted an article justifying discrimination against individual Black people on the grounds that “Black people as a race have been tainted by their lousy upbringing and single-parent homes.” No one would tolerate such a Tweet and no one should tolerate this.
3
33
577
Meta Reality Labs is gonna cost as much as the Apollo Space Program, or about 10 Manhattan Projects.
🫠
132
820
3,364
#ECCV2022 Check our new work (MalleConv) about significantly speedup image denoising on tomorrow‘s poster session！This is my intern project at , mentored by Tianfan Xue, , , . And will be presenting in-person!
1
4
12
Show this thread
and also ours (zhiwenfan.github.io/INS/). If you are a big fan of stylization, don’t forget to see all above fancy works!
Quote Tweet
There have been a number of really nice papers on 3D style transfer recently, including CLIP-based Neural Neighbor Style Transfer for 3D Assets:
granskog.xyz/clip-nnfm-for-
and UPST-NeRF:
semchan.github.io/UPST_NeRF/
and SNeRF:
research.facebook.com/publications/s
Show this thread
2
5
👍👍
Quote Tweet
We at @MITIBMLab, are currently looking to hire PhD candidates for 2023 summer internship, to work on efficient training and inference of large language (and/or vision) models. Please DM or send me an email if you are interested.
This looks surprisingly good…
Quote Tweet
Need help imagining what your next animated character could look like? Make-A-Video can help with that 
#MetaAIMakes
#MetaAIMakes
0:14
57.3K views
Some ML research areas are extremely crowded. Working there is a race. But it's not about "how to benefit science" but "who gets the credit." This is the furthest from being a scientist I can imagine. How did we get here?
111
160
1,317
If you set your research goal as to publish in the next conference, then it is highly probable you will never amount to anything significant.
20
61
568
Isn't that just a simple “flow estimation + pixel warping”？
Quote Tweet
Used Dall-e2 to finish background of the video. AI + 10 minutes in Adobe AE and here’s great result!
Show this thread
0:04
78.8K views
Check our latest research on efficient image processing @ ECCV！ twitter.com/BartWronsk/sta
This Tweet is unavailable.
4
An interactive map of the current average price for a gallon of regular unleaded gasoline in each state.
603
878
2,532
Curious how MobileNeRFs would look in VR? I've hacked WebXR support into the online demo viewer: mobilenerf-webxr.fern.solutions (requires PCVR for best experience) #threejs #webxr
Make sure to also checkout the original: mobile-nerf.github.io
10
92
485
“I didnt” means I have concerns that reviewer will attack us using this evidence🤣
Show this thread
I remember that I've discussed with a senior collaborator in industry about this while he trusted reviewer will appreciate us showing failure cases but I didnt. Not sure if that happened to most industry researchers （surely they care more than academic researchers due to product
Quote Tweet
I like to show the failure cases of a method when I write a paper. Unfortunately, reviewers usually bring up the failure cases as a major weakness. This will probably kill at least one of my #NeurIPS papers. I can see why cherry picking has become the norm in some fields.
1
Show this thread
seems every academic researcher has met this complain🤣. DL generalization is so poor
Quote Tweet
Here's an example of how saliency maps can be useful. I recently met with a lab at NIH that builds automatic diagnosis models from MRI images. They built a conv net that performed well on their validation images, but it failed when applied to images from new sources...
twitter.com/zacharylipton/…
twitter.com/zacharylipton/…Show this thread
The Common Objects in 3D v2 #CO3Dv2 dataset is the next version of the first real-life dataset enabling large-scale 3D reconstruction of object categories. Together w/ the new version, we're also pleased to announce the CO3D challenge happening at #ECCV22! eval.ai/web/challenges
3
69
195
Show this thread
Amazing！
Quote Tweet
1
You use GPUs everyday, but do you (actually) know how they work?
GPU-Puzzles (v0.1) - 14 short puzzles in Python with a visual debugger. No background required. Do puzzles, learn CUDA.
Link: github.com/srush/GPU-Puzz
26
547
3,407
Show this thread
NASA: We took high-res pictures of galaxies over 13 billions years ago.
Computer Vision: We reduced ImageNet classification errors by 0.1%!
5
16
219
Can not agree more.
Is one month “from-zero-to-arxiv research” really research？CV community is going worse
Quote Tweet
We community should do something. 1 Add to vision conferences a mechanism to detect plagiarism. 2. Ban ArXiv and slow down science to reduce everyone’s publication pressure. I’m too junior to make them happen, but happy to be a co-signer if someone could propose them to PAMI-TC. twitter.com/CVPR/status/15…
Finding it daunting to complete a decent voxel drawing in 99 lines of Python code? Not to worry. Yuanming Hu, the creator of #TaichiLang, shares how to build a voxel autumn theme from zero to hero in his latest post: bit.ly/3Lkr4ol. Take a look!
8
51
Who can explain all these "celebrated" papers claiming MLP<CNN<ViT<CNN<MLP... Aren't people having better things to do? We need a theory and a framework that unifies all these instances, not fighting over trivia...
10
25
256
Here's this year's "call for interns" for our tech area in Google Research. We've organized into a ~100 person team under David Salesin and Bill Freeman, with a focus on anything and everything involving photography, media, creativity, and HCI. My sub-team works on NeRF.
4
24
174
Show this thread