Lightnews — Scholar-powered news

Evangelos Kazakos

@ekazakos.bsky.social

690 followers 460 following 640 posts

Postdoctoral researcher @ CIIRC, CTU, Prague working in vision & language. Also robotics noob. PhD from University of Bristol. Ex. Samsung Research (SAIC-C). I love coffee and plants. And socks.

Posts Media Videos Starter Packs

Pinned

Evangelos Kazakos @ekazakos.bsky.social · Jun 26

Our work, GROVE, has been accepted to ICCV 2025! 🎉 This is collab. w. Cordelia Schmid & @josef-sivic.bsky.social.

We will release code, models and datasets within next 2 weeks.

We are also working on a search demo for the proposed datasets with user prompts!

I hope to see you all in Honolulu!

Evangelos Kazakos @ekazakos.bsky.social · Apr 16

🚨🚨📢📢 PAPER ALERT 📢📢🚨🚨

Excited to announce our new work: "Large-scale Pre-training for Grounded Video Caption Generation" with Cordelia Schmid & @josef-sivic.bsky.social.
Paper: arxiv.org/abs/2503.10781
Project: ekazakos.github.io/grounded_vid...
Code (coming soon): github.com/ekazakos/grove 1/7

Large-scale Pre-training for Grounded Video Caption Generation

We propose a novel approach for captioning and object grounding in video, where the objects in the caption are grounded in the video via temporally dense bounding boxes. We introduce the following con...

arxiv.org

1 4 23

Reposted by Evangelos Kazakos

Yoshitomo Matsubara @yoshitomo-matsubara.net · 15h

Did you open-source your #ICCV2025 works?

As a PyTorch Ambassador, I would like to write an article to introduce open-sourced ICCV 2025 works (including workshops and demos) for promoting open-source/science + PyTorch

If interested, share your work via the form in my reply 👇

2 5 5

Evangelos Kazakos @ekazakos.bsky.social · 10h

🙌

Reposted by Evangelos Kazakos

Curated Data Workshop @ ICCV '25 @curated-data.bsky.social · 13h

Our schedule is live!

curateddata.github.io/schedule.html

We have posters, oral presentations, and invited talks from @zhuangl.bsky.social, @sarameghanbeery.bsky.social, @andrewowens.bsky.social, @phillipisola.bsky.social, and Alyosha Efros!

See you in Hawaii!

#ICCV2025 @iccv.bsky.social

Curated Data for Efficient Learning

The ICCV 2025 Workshop on Curated Data for Efficient Learning

curateddata.github.io

1 6

Reposted by Evangelos Kazakos

Sara Beery @sarameghanbeery.bsky.social · 13h

This is going to be fun! I'll be talking about our recent work on dataset subselection for specialization instead of generalization - how do we build optimal models for specific end users given a big pool of generalist data?

Curated Data Workshop @ ICCV '25 @curated-data.bsky.social · 13h

Curated Data for Efficient Learning

The ICCV 2025 Workshop on Curated Data for Efficient Learning

curateddata.github.io

1 2 11

Evangelos Kazakos @ekazakos.bsky.social · 1d

Ok, the computer vision classification convinced me to at least try it. I will give it a go! And thanks!

1 1

Evangelos Kazakos @ekazakos.bsky.social · 1d

I want to make a feed for the upcoming ICCV 2025. Previously, I've built a computer vision (bsky.app/profile/did:...) using @skyfeed.app. Do you reckon @graze.social is better? Thanks!

1 1

Reposted by Evangelos Kazakos

Kristina Šekrst @epiphenomenal.bsky.social · 2d

Accurate.

#academia #academy #philsky #philosophysky #philosophy #science

3 79 380

Evangelos Kazakos @ekazakos.bsky.social · 1d

😂

Reposted by Evangelos Kazakos

humphreyshi.bsky.social @humphreyshi.bsky.social · 3d

CVPR 2026 Workshop Call is officially open! 🎉

📅 Deadline: Nov 3, 2025
cvpr.thecvf.com/Conferences/...

We expect to host over 100 workshops this year — and this is where the next breakthroughs begin. 🚀

With Mei Chen,Erik Blasch,Brian Clipp,Sharon Huang,Andreas Savakis @cvprconference.bsky.social

7 14

Evangelos Kazakos @ekazakos.bsky.social · 2d

As Jitendra Malik is saying "Robotics is too important to leave it to the robotics people".

Christian Wolf @chriswolfvision.bsky.social · 3d

Computer vision will not only save robotics, it will save AI 😆

Chaz Firestone @chazfirestone.bsky.social · 3d

This is a big one! A 4-year writing project over many timezones, arguing for a reimagining of the influential "core knowledge" thesis.

Led by @daweibai.bsky.social, we argue that much of our innate knowledge of the world is not "conceptual" in nature, but rather wired into perceptual processing. 👇

1 1 4

Reposted by Evangelos Kazakos

Johan Edstedt @parskatt.bsky.social · 3d

All papers I was assigned for 3DV were significantly higher quality than the papers I was assigned for ICCV.

2 1 7

Reposted by Evangelos Kazakos

dimosk.bsky.social @dimosk.bsky.social · 3d

PhD Position open, through the ELLIS PhD program (ellis.eu/news/ellis-p...), at the Computer Vision Centre (www.cvc.uab.es), Barcelona! Linked to #ELLIOT project (www.elliot-ai.eu) a major European initiative to develop #open, #trustworthy and #multimodal #FoundationModels. @elliot-eu.bsky.social

ELLIS PhD Program: Call for Applications 2025

The ELLIS mission is to create a diverse European network that promotes research excellence and advances breakthroughs in AI, as well as a pan-European PhD program to educate the next generation of AI...

ellis.eu

2 7

Evangelos Kazakos @ekazakos.bsky.social · 2d

I mean that probably explains why they are always angry... 😂

1 1

Evangelos Kazakos @ekazakos.bsky.social · 2d

They seem like a couple that just had a fight and can't tolerate each other.

1 2

Reposted by Evangelos Kazakos

Alex Spahn 🌋🌪️☄️ @spahn711.bsky.social · 2d

One year ago today, a spectacular aurora display graced the night sky across much of the globe. This was my view in *Oklahoma.* What a beautiful world we are fortunate to live on. #photography

Pink, green, and orange aurora pillars with a SAR arc and moon-lit clouds over relatively still waters of a large lake.

The andromeda galaxy between a red SAR arc and the Milky Way.

Numerous pinkish aurora pillars with a bluish-purple sky and spotty clouds over relatively still waters of a large lake.

Pink, green, and orange aurora pillars with a bluish-purple sky over relatively still waters of a large lake.

13 130 630

Reposted by Evangelos Kazakos

Vicky Kalogeiton @vickykalogeiton.bsky.social · 4d

Very proud of our recent work, kudos to the team! Read @davidpicard.bsky.social’s excellent post for more details or the paper arxiv.org/pdf/2502.21318

6 17

Evangelos Kazakos @ekazakos.bsky.social · 3d

Congratulations Dr. Edstedt! 🎉

Reposted by Evangelos Kazakos

Andreas Geiger @andreasgeiger.bsky.social · 3d

Personal programs for ICCV 2025 are now available at:
www.scholar-inbox.com/conference/i...

6 22

Reposted by Evangelos Kazakos

Dmytro Mishkin @ducha-aiki.bsky.social · 4d

For those going to @iccv.bsky.social, welcome to our RANSAC tutorial on October 2025 with
- Daniel Barath
- @ericbrachmann.bsky.social
- Viktor Larsson
- Jiri Matas
- and me
danini.github.io/ransac-2025-...
#ICCV2025

1 6 20

Evangelos Kazakos @ekazakos.bsky.social · 6d

To keep the tradition, the lineup is 🔥🔥🔥

Giorgos Tolias @gtolias.bsky.social · 6d

The Visual Recognition Group at CTU in Prague organizes the 50th Pattern Recognition and Computer Vision Colloquium with
Torsten Sattler, Paul-Edouard Sarlin, Vicky Kalogeiton, Spyros Gidaris, Anna Kukleva, and Lukas Neumann.
On Thursday Oct 9, 11:00-17:00.

cmp.felk.cvut.cz/colloquium/

3 11

Reposted by Evangelos Kazakos

Shahab Bakhtiari @shahabbakht.bsky.social · 6d

Interesting paper suggesting a mechanism for why in-context learning happens in LLMs.

They show that LLMs implicitly apply an internal low-rank weight update adjusted by the context. It’s cheap (due to the low-rank) but effective for adapting the model’s behavior.

#MLSky

arxiv.org/abs/2507.16003

Learning without training: The implicit dynamics of in-context learning

One of the most striking features of Large Language Models (LLM) is their ability to learn in context. Namely at inference time an LLM is able to learn new patterns without any additional weight updat...

arxiv.org

1 19 60

Evangelos Kazakos @ekazakos.bsky.social · 7d

I've added a rule to add posts from this account always in the feed so no need for the tag 😉 But thanks a lot for caring about it mate! ❤️

1 1

Evangelos Kazakos @ekazakos.bsky.social · 7d

#skyvision

Reposted by Evangelos Kazakos

Sung Kim @sungkim.bsky.social · 9d

ModernVBERT

A vision-language encoder that matches the performance of models 10× its size on visual retrieval tasks!

📄 Paper: arxiv.org/abs/2510.01149
🌐 Blog: huggingface.co/blog/paultlt...
👨‍🍳 Finetuning Cookbook: colab.research.google.com/drive/1bT5LW...
🤗 Models: huggingface.co/ModernVBERT

1 2 14

Evangelos Kazakos @ekazakos.bsky.social · 8d

If you install ROCm version of PyTorch for AMD, you can of course use it without any changes in your code and with very negligible differences in performance (without this meaning that CUDA performs better).