Evening reading. I’ve been looking forward to reading this paper. @chriswolfvision.bsky.social @weinzaepfelp.bsky.social
Kosta Derpanis
@csprofkgd.bsky.social
#CS Associate Prof York University, #ComputerVision Researcher, VectorInst Faculty Affiliate, TPAMI AE, ELLIS4Europe Member, #ECCV2026 Publicity Chair 📍Toronto 🇨🇦 🔗 csprofkgd.github.io 🗓️ Joined Nov 2024
When I did my PhD, ICPR was this big conference and CVPR was a much smaller high-quality venue. The trends in sizes were clearly different.
Taking my talk on the road today. Visiting ETH Zürich, hosted by Siyu Tang 🤗
For years, I've heard from my fellow computer scientists that "the brain is a computer" or "we can simulate brains in a computer." I think both are wrong, and, now that we're hearing them from corporate tech leaders too, it's no longer just an academic topic. 🧵 aaronhertzmann.com/2026/07/13/b...
The Brain Is Not Just A Computer, and Cannot Be Simulated by One
Digital computers perform fixed logical operations on clock cycles; brains are biological systems of staggering complexity.
aaronhertzmann.com
Next stop: Zurich for the final leg of my sabbatical, through the end of July 🤗
The #ECCV2026 workshop page with links is now up! eccv.ecva.net/Conferences/...
2026 Workshops
eccv.ecva.net
Introducing my new #DeepLearning course page 🤗 The page is now live and evolving over the coming weeks. Check it out: csprofkgd.github.io/deep-learnin... Feedback is welcome!
Next stop: home for the week ❤️ Then back to Zurich to wrap up the final stretch of my sabbatical.
#KostasThoughts: The #computervision community is behind on review transparency. Reviews and meta-reviews should be made public for accepted papers and, if authors opt in, for rejected papers as well. Transparency can be a powerful mechanism for improving review and meta-review quality. 1/2
The face of academic research in 2026: always GPU-constrained, and now token-constrained too. 🤦
#ECCV2026 decisions have been released! The final reviews and meta-reviews are not yet available on OpenReview; they will be released in the coming days.
#ECCV2026 decision notifications will be sent via email by June 17th anywhere on Earth.
Today was my last day at the Samsung AI Centre Toronto. It’s been an AWESOME 8-year run. I’ve learned a tremendous amount, worked with incredible people, and built lasting friendships along the way. 1/2
Inspired by @jbhuang0604.bsky.social's Awesome-Tips webpage and his encouragement, I put together a webpage collecting #KostasThoughts posts from over the years. Hope you find them useful! #KostasThoughts: csprofkgd.github.io/kostas-thoug...
#KostasThoughts
YouTube video by CSProfKGD
youtube.com
Last few days to submit your artwork to @eccv.bsky.social 👀 Submissions deadline 14th June: eccv.ecva.net/Conferences/... With Ioannis Siglidis #ECCV2026 📷 from #ECCV2018 gallery @annaridler.bsky.social @philippschmitt.bsky.social Masaru Mizuochi #creativeAI #AIart
First full draft of my 🆕 lecture on (vanilla) policy gradients done 😅 TODO: Variance reduction methods (temporal structure and baselines).
Vanilla Policy Gradients
YouTube video by CSProfKGD
youtu.be
Back to the lecture prep grind. VERY early version of my lecture on Policy Gradients. With #computervision and #robotics research appearing to converge, at least based on #CVPR2026 and my European tour conversations, it feels like good timing 😉
The videos from the “Frontiers of Embodied AI” meetup at ETHZ from a few weeks back are now available. Speakers: Jitendra Malik, Vladlen Koltun, Yann LeCun, and Shuran Song Hosted by Marc Pollefeys YouTube playlist: youtube.com/playlist?lis...
When co-authors independently write different parts of a paper 😬
#KostasThoughts My lecture motto: Your eyeballs are mine. One idea at a time. The focus of the slide should always match the focus of the speech.
Back to the lecture prep grind. VERY early version of my lecture on Policy Gradients. With #computervision and #robotics research appearing to converge, at least based on #CVPR2026 and my European tour conversations, it feels like good timing 😉