Computer Vision

1 curation · 69 membres · fenêtre de 168 h, demi-vie 6 h, seuil 2 relayeurs

Fenêtre 6 h 12 h 24 h 2 j 3 j 7 j
Montrer Les deux Liens 1 Tweets 7
Seuil 2 relayeurs 3 relayeurs 4 relayeurs 5 relayeurs 6 relayeurs Tout remettre
1liens
7tweets
2657posts lus
57auteurs
403liens en base
1435 hde recul

Les liens 1

Classés par auteurs distincts, jamais par audience.

1

Solv3R: 3D Looped Transformer as a Multi-view Solver

One learned solver for offline, incremental, online, and any-order 3D reconstruction.

2 relayeurs à 2 · il y a 6 j ubc-vision.github.io
Qui l'a relayé
Dmytro Mishkin @ducha_aiki · 03/11 11:06 Marrying classical CV and Deep Learning. I do things, which work, rather than being novel, but not working.
📢📢📢 Introducing 𝐒𝐨𝐥𝐯𝟑𝐑 Can geometric foundation models move beyond fixed-set prediction and keep solving like classical SfM? We turn a looped Transformer into a reusable 𝐥𝐞𝐚𝐫𝐧𝐞𝐝 𝐠𝐞𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐬𝐨𝐥𝐯𝐞𝐫 🌐 Project: 🧵👇
Kosta Derpanis @CSProfKGD · 03/11 04:49 #CS Associate Prof @YorkUniversity #ComputerVision Researcher, @VectorInst Affiliate @UofT status only, @ELLISforEurope Member, X-social media chair 4all vision
📢📢📢 Introducing 𝐒𝐨𝐥𝐯𝟑𝐑 Can geometric foundation models move beyond fixed-set prediction and keep solving like classical SfM? We turn a looped Transformer into a reusable 𝐥𝐞𝐚𝐫𝐧𝐞𝐝 𝐠𝐞𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐬𝐨𝐥𝐯𝐞𝐫 🌐 Project: 🧵👇

Les tweets 7

Axe indépendant : repris sans regarder les liens.

1
Richard Sutton @RichardSSutton · 08/11 05:00
Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHO
3 relayeurs à 3 · il y a 22 h Voir sur X
Qui l'a repris
Satya Mallick” martin_casado Kosta Derpanis
Ce qu'ils en ont dit
Satya Mallick @LearnOpenCV a cité · 08/11 05:00
That’s what I call a bitter sense of humor.
2
Michael Black @Michael_J_Black · 06/11 19:44
The dogma of the last decade in computer vision is that all classical algorithms should be replaced by neural networks. This has been very productive but, counterintuitively, with the latest frontier models, I find myself sometimes turning back to classical methods. If you have
3 relayeurs à 3 · il y a 2 j Voir sur X
Qui l'a repris
Matthias Niessner Satya Mallick” Kosta Derpanis
Ce qu'ils en ont dit
Satya Mallick @LearnOpenCV a cité · 06/11 17:41
A lot of algorithms, especially in geometry are difficult to code and very delicate. Now, thanks to frontier models we can try those ideas out. Not everything needs to be solved using a big hammer of neural networks.
3
Justin Johnson @jcjohnss · 06/11 02:21
We've released a demo of real-time world generation powered by our new Atlas model! This is a spiritual successor to our RTFM demo from last year. It's a vision for the future of interactive media might look like, with frames streamed directly from a powerful generative model
2 relayeurs à 2 · il y a 3 j Voir sur X
Qui l'a repris
martin_casado Kosta Derpanis
4
Yann LeCun @ylecun · 05/11 02:01
@Arjunjain Plot twist: I was Vapnik's Department Head between 1996 and 2002, and made sure he had the space, resources, and collaborators to do his best work. To people who think I'm dogmatic: I famously disagreed with Vapnik about the best way forward in ML, but did what I could to
2 relayeurs à 2 · il y a 4 j Voir sur X
Qui l'a repris
Kosta Derpanis Satya Mallick
5
Andrea Tagliasacchi @taiyasaki · 03/11 11:06
📢📢📢 Introducing 𝐒𝐨𝐥𝐯𝟑𝐑 Can geometric foundation models move beyond fixed-set prediction and keep solving like classical SfM? We turn a looped Transformer into a reusable 𝐥𝐞𝐚𝐫𝐧𝐞𝐝 𝐠𝐞𝐨𝐦𝐞𝐭𝐫𝐢𝐜 𝐬𝐨𝐥𝐯𝐞𝐫 🌐 Project: 🧵👇
ubc-vision.github.io Solv3R: 3D Looped Transformer as a Multi-view Solver One learned solver for offline, incremental, online, and any-order 3D reconstruction.
3 relayeurs à 3 · il y a 6 j Voir sur X
Qui l'a repris
Dmytro Mishkin Kosta Derpanis F Frank Dellaert”
Ce qu'ils en ont dit
F
Frank Dellaert @fdellaert a cité · 03/11 03:41
Solv3R looks amazing!
6
R
Rikiya Takehi @rikiyatakehi · 02/11 23:51
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming:
3 relayeurs à 3 · il y a 6 j Voir sur X
Qui l'a repris
Yoshitomo Matsubara” A Aviad Dahan” Douglas Gray”
Ce qu'ils en ont dit
Yoshitomo Matsubara @yoshitomo_cs a cité · 02/11 23:51
Sounds like a surprising result Is it possible that GPT-6 Astra is aware of the target task and have access to some fine-tuned Qwen weights in some way (say 🤗) to cheat? Some of the replies imply that it won't generalize well
A
Aviad Dahan @AviadDahan a cité · 02/11 22:33
it's like reverse engineering assembly code but for neural networks
Douglas Gray @de3ug a cité · 02/11 20:50
Franticly updating priors about what the future of ML research looks like… Spoiler: “Just Ask”
7
O
Oliver Hahn @olvr_hhn · 02/11 23:07
📢 LucidLyra: Feed-Forward Generative 3D Scene Reconstruction and Joint Understanding via Multi-Task Distillation! Generate a queryable 3D Gaussian scene from a single image! Our feed-forward decoder jointly predicts geometry, appearance, open-vocabulary semantics, and object
3 relayeurs à 3 · il y a 6 j Voir sur X
Qui l'a repris
L Laura Leal-Taixe Kosta Derpanis Stefan Roth