@ekazakos
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. ๐ค๊ฉ๐๐ฐ๐ค
A paid summer internship is available in Microsoft Research NYC for a *PhD student later in their degree*, with experience designing human questionnaires/experiments, collecting data on prolific, running statistical analysis, and writing papers. If you know exceptional candidates please ping me. ๐
They find that pretrained Vision-Language-Action (VLA) models are surprisingly resistant to forgetting!
Zero forgetting, or even positive backward transfer, is possible with simple experience replay.
Let me introduce our new paper: Multimodal Large Language Models as Image Classifiers
โ Multimodal LLMs are increasingly used for visual tasks, but evaluating their image classification ability has produced conflicting conclusions.
Link: arxiv.org/html/2603.06...
More than 30 OpenAI and Google DeepMind employees signed onto a statement supporting Anthropic's lawsuit against the Defense Department after the agency labeled the AI firm a supply chain risk, according to court filings.
@atabb.bsky.social @teafrancis.bsky.social
Next week Iโll be in Bristol as a PhD viva examiner and to give a talk, hosted by @dimadamen.bsky.social. Looking forward to it ๐ค
We are aware of recent reports regarding targeted phishing attacks that have resulted in account takeovers of some Signal users, including government officials and journalists. We take this very seriously. 1/7
1/n #CVPR2026 Accepted Paper๐
๐จ๐๐ ๐ ๐ญ๐๐ ๐ฌ๐๐๐๐๐๐๐ ๐ฌ๐๐๐๐๐ ๐๐ ๐ฉ๐๐๐
๐๐ ๐๐๐ ๐บ๐๐๐๐๐๐๐๐๐๐ ๐ฎ๐๐ ๐๐ ๐ถ๐๐๐-๐ฝ๐๐๐๐๐๐๐๐๐ ๐บ๐๐๐๐๐๐๐๐๐๐๐?
๐น๐๐๐๐๐๐๐ ๐๐๐
๐บ๐๐๐๐๐๐ (๐น๐ต๐บ) answers this question.
Paper/code at the end๐๐ผ
fun fact: the most concrete specification of a program you can formulate is called CODE.
New #Survey Certification:
Open Technical Problems in Open-Weight AI Model Risk Management
Stephen Casper, Kyle O'Brien, Shayne Longpre et al.
https://openreview.net/forum?id=8QyGLnFkzc
#risks #weights #risk
SiLVR: A Simple Language-based Video Reasoning Framework
Ce Zhang, Yan-Bo Lin, Ziyang Wang, Mohit Bansal, Gedas Bertasius
Action editor: Anurag Arnab
https://openreview.net/forum?id=mQZbh9Zlbw
#multimodal #subtitles #captions
Itโs already happening. Iran attacked a desalination plant in Bahrain.
This war has to end immediately. If it turns into a water war we will be dealing with a whole different level of a humanitarian catastrophe across the region.
aje.news/64yl8m?updat...
ฮงฯฮฟฮฝฮนฮฑ ฯฮฟฮปฮปฮฑ!
Please please read this from the archbishop of Chicago
www.archchicago.org/statement/-/...
Caitlin KALINOWSKI over X I resigned from OpenAl. I care deeply about the Robotics team and the work we built together. This wasn't an easy call. Al has an important role in national security. But surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got. This was about principle, not people. I have deep respect for Sam and the team, and I'm proud of what we built together.
OpenAI head of robotics just resigned over company deal with the Pentagon sayingโฆ
โSurveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they gotโ
๐
ฮฮณฯ ฯฮฑฯ ฯฮฟฯ ฯฮตฯฮผฮฌฯฮบฮตฯ ฮณฮนฮฑ ฮฝฮฑ ฮฒฮณฯ ฮฑฯฮฟ ฯฮฟ ฯฯฮนฯฮน ฮฝฮฑ ฮพฮตฯฮบฮฑฯฯ ๐
"I didn't see a single building standing the entire time I was in Gaza. I didn't see one building intact or standing.
Everything was flattened. And when I mean flattened, it was hard to even tell that they were buildings there before."
www.instagram.com/reels/DVMK-u...
the harder the industry invests on pushing narratives of AI as (only) positive, inevitable, and inherently good for society/business, the more any criticism of this narrative becomes โtoo radicalโ, โunworkableโ and โunrealisticโ
I guess here is the perfect chance to advertise our workshop: aiforpeaceworkshop.github.io.
Iโd appreciate a lot if you could take a look, thanks!
Yep, we fully agree here. ๐
I was saying today exactly the same. A bunch of **dumb** motherfuckers are destroying the whole world.
Totally reminds me of this movie.
"Anthropic has much more in common with the Department of War than we have differences."
๐ข๐ We have open PostDoc positions in Computer Vision & ML at @tuda.bsky.social and @hessianai.bsky.social within the Reasonable AI Cluster of Excellence โ supervised by @stefanroth.bsky.social, @simoneschaub.bsky.social and many others!
Apply here: www.career.tu-darmstadt.de/tu-darmstadt...
I love this thread. โค๏ธ๐
Hilde is, as often, spot on.
LLMs can speed-up things, like the tedious work to make a pretty plot in TikZ, but this is not research. It's next to research: a necessary task, but not the thing-in-itself. For those, it's amazing. But for the part that requires critical thinking, not much indeed.
@hildekuehne.bsky.social: "Take off the day ... After proofreading a stack of papers, I realized that there is just no way to speed up research with LLMs ... you can just speed up random nonsense. Problem is that only a few people are able to tell the difference."
Enjoying the For You feed? Give it a like โก to help more people discover it: bsky.app/profile/did:...
The more people use it -> the more feedback we get -> the better we can make it for you.
