How can one get the tweets of AI research folks but not the tweets of AI influencers? Asking for a friend.
Christian Wolf (🦋🦋🦋)
5,840 posts
Christian Wolf (🦋🦋🦋)
@chriswolfvision
Principal Scientist, @NaverLabsEurope, Lead of Spatial AI team. AI for Robotics. Feedback: admonymous.co/chriswolfvision
- That's a nice way of writing equations (I sometimes do this in lectures). From ICLR 2021 submission ("An attention free transformer"), openreview.net/forum?id=pW--c…
- I passed this image through a pretrained ResNet 101 with @PyTorch and it predicts: Tiger shark 23% Hammerhead 21% Great white shark 16% Gar, garfish 11% Sturgeon 3% My conclusion : shape > [ texture + mountain context] => reassuring in some way. (Image from @SolTight)
- This is on of the best papers I have recently read, congrats for this excellent work. "NNs do not have an inherent “simplicity bias”. This property depends on components such as ReLUs, residual connections, and layer normalizations". This also applies to transformers.
Replying to @francoisfleuretWhatever signal has the same shape as that preferred by your architecture (shameless plug: Neural Redshift arxiv.org/abs/2403.02241). - Peak AI happened finally ... #NeurIPS2021 "We received 9,052 full paper submissions this year, a slight decrease compared to last year."
- I trained a CNN w/o pooling on MNIST and produced a colored visualization of the feature maps you get when you translate a digit on an input canvas. This illustrates the equivariance property of convolutions. More info (why convolutions?) in a blog post: medium.com/@chriswolfvisi…
GIF - Just created a new Figure for a survey paper, I am quite satisfied.
- I love this figure showing the difference of manifolds learned by GANs and VAEs on a 1D toy example. The VAE learns the mean, the GAN passes through the data with higher fidelity but does not cover the full domain. From: Plumerault et al. ICPR 2020 arxiv.org/abs/2012.11551
- This paper had been rejected 4x at conferences and now got the #cvpr2024 best student paper runner up reward! arxiv.org/abs/2212.06872
- I can fix computers only if they have a tape of infinite length.
- Integral neural networks look like a fantastic innovation: continuous representations along filter and channel dimension, prune w/o finetuning. It is rare to discover a paper when the #CVPR proceedings are published, this paper was not on arxiv before.. openaccess.thecvf.com/content/CVPR20…
- This new Biography of John von Neumann was an incredible read. I am not sure he was really human... The book is extremely well written and puts all his ideas and work into its historical context, eg the work of Turing and Gödel, and follow-up work, eg Conway etc.
- A thread on similarities in the ways the field addressed two seemingly different problems, namely A) recent work on transformers / self-attention addressing their quadratic complexity, and B) work decoupling capacity and memory size in recurrent neural networks. 👇 1/19
- < 2018: we have tried instance norm, batch norm, layer norm ... 2018: let's normalize over the missing dim, channel groups => ECCV 2018 best paper < 2023: we applied the Fourier transform over 1d signals, images, time .... 2023: let's FFT channels => ICLR ratings of 8, 8, 10, 8.
How do you get the highest ICLR ‘24 review scores (8, 8, 10, 8) ? You revitalize your old course notes for Image Processing 101.
