@VIPriorsICCV

Visual Inductive Priors for Data-Efficient Deep Learning Workshop @ ICCV 2023 in Paris. Tweets by @rjbruin.

vipriors.github.io
Joined February 2020
  • Tweets71
  • Following112
  • Followers237
8Media
VIPriors Workshop @ ICCV 2023 retweeted
VIPriors Workshop @ ICCV 2023@VIPriorsICCV13 Jul 2023
Paper rejected from #ICCV2023? Submit to the VIPriors workshop and maybe you still get to go to Paris! 💬 Topic: inductive biases & data efficiency 📄 Format: same as ICCV 🗓️ Deadline: July 16th 🌐 vipriors.githhub.io/call-for…
1
1
1
6,312
VIPriors Workshop @ ICCV 2023 retweeted
Efstratios Gavves@egavves28 Jun 2023
An excellent workshop on a very timely moment? Question for the curious: what is the role of priors in the era of foundation models and the GPTs? An important direction to explore, and hopefully an excellent lineup of speakers can say one thing or two!
VIPriors Workshop @ ICCV 2023@VIPriorsICCV26 Jun 2023
Working on prior knowledge in Computer Vision? Consider submitting to our workshop @ #ICCV2023! Topics include: geometric deep learning, data augmentation, biases that benefit data efficiency, and more! 📅 Deadline: July 16th 🌐 CfP: vipriors.github.io/call-for-…
2
8
1,148
VIPriors Workshop @ ICCV 2023 retweeted
Andy Keller@t_andy_keller11 Oct 2021
If you're at ICCV today, come check out our oral presentation at the @VIPriorsICCV workshop! Predictive Coding with Topographic VAEs: openaccess.thecvf.com/conten… We show how the 'capsule roll' can be used directly for forward prediction, greatly improving transformation modeling!
Forward prediction log-likelihood vs. future time offset δt. We see that the PCTVAE has consistently high likelihood for sequence elements into the future whereas the likelihood of the TVAE model drops off rapidly. Shading denotes ± 1 std.

ALT Forward prediction log-likelihood vs. future time offset δt. We see that the PCTVAE has consistently high likelihood for sequence elements into the future whereas the likelihood of the TVAE model drops off rapidly. Shading denotes ± 1 std.

1
5
1
27
VIPriors Workshop @ ICCV 2023 retweeted
Björn Barz@Callidior3 Sep 2021
In our recent @VIPriorsICCV paper, we introduce a benchmark for data-efficient image classification and re-evaluate 8 state-of-the-art methods, tuning all of them carefully on held-out data. Surprisingly, only one method outperforms the baseline slightly. arxiv.org/abs/2108.13122
2
1
1
7
VIPriors Workshop @ ICCV 2023 retweeted
Ivan Sosnovik@isosnovik1 Sep 2021
I'm happy to announce that 2 our joint works with brilliant @artemmoskalev and @arnoldsmeulders got accepted to @VIPriorsICCV Relational Prior for Multi-Object Tracking (Oral) openreview.net/forum?id=1MZn… How to Transform Kernels for Scale-Convolutions openreview.net/forum?id=rTpT…
1
3
20
VIPriors Workshop @ ICCV 2023 retweeted
ICCV2021@ICCV_202128 Jul 2021
Now that the papers for the main #iccv2021 conference have been selected, please turn your attention to the impressive lineup of exciting workshops. iccv2021.thecvf.com/node/44 There are over 80 workshops covering a wide range of specialized topics of interest to the vision community
12
4
56
Load more