Pavol Harar

3/7 🏎 We developed FakET as a surrogate model mimicking the behavior of the physics-based transmission electron microscope simulator SHREC but FakET generates the data 750x faster and uses 33x less memory 💪

Finally, my postdoc work is published in Cell Structure! 🎉 Grateful for the chance to apply deep learning to cryo-ET and learn from @HaselbachLab about structural biology🦠. Huge thanks to Lukas Herrmann for coding help and @GrohsPhilipp for my position and all the GPUs.

This week the 7th Austrian Cryo-EM Symposium happened & we were thrilled to once again welcome so many participants on our campus! Many thanks to the organizers @IMPvienna, Vienna BioCenter Core Facilities (VBCF) and @ISTAustria for a great event! #viennabiocenter

@aks1812 @glouppe Bold of you all assuming it is a human.

@glouppe Wao....lol...then it seems mischievous to give such a review

This is a first... Reviewer #2 asking us to compare against the very paper we are submitting and that he is (supposed to be) reviewing!? This is all just a farce... #ICLR2025 🙃

@bengeliscious @svenklumpe @biozentrum @svenklumpe for the win!

The super talented @svenklumpe visited the @biozentrum to teach our students about #cryoET workflows, cryo lift-out, and his very cool research into retrotransposon particles. Thanks Sven!

@GrohsPhilipp @ISTAustria @HaselbachLab Thanks, Philipp! It’ll be hard to top the fun I had in your group. By the way, that's my favorite cup from Univie, had to come with me.😇

@HararPavol @ISTAustria @HaselbachLab Congratulations Pavol! I hope you'll have a great time! The coffee looks good :)

Thanks, Alicia! I'm very happy to join your group at @ISTAustria 🎉. At the same time, I need to thank my mentors @GrohsPhilipp , @HaselbachLab, Monika Dörfler, Jiri Mekyska, Lukas Vrabel, and all my great former colleagues. Without them I would not make it this far.

@jankosinski @ERC_Research Yeah, @jankosinski, there is no better feeling then deploying a new beast like this 🤩 I remember it well.

Two 🔥 computational postdoc positions 🔥are available in our lab, funded by @ERC_Research ! embl.org/jobs/position/… and embl.org/jobs/position/…. Join us and unleash the power of these just-delivered beasts 👇And please repost to help us!

In the spirit of Obama's recent viral post about getting shi<ehm>stuff done, I just submitted two major updates of my papers on arXiv. Not every day you push through multiple submissions at once, even if it's "just" updates. Let's see if I manage to get another one next week.

The next Vienna Deep Learning Meetup takes place on June 19, 2024 at 18:30 hosted by Bosch. We look forward to seeing you there meetup.com/vienna-deep-le…

I’ve joined the organization team of the Vienna Deep Learning Meetup! My first initiative is merging the Vienna Computer Vision Meetup with VDLM to unify the Viennese AI community. Let's innovate together! 🚀 #VDLM #AI #DeepLearning #ComputerVision meetup.com/vienna-deep-le…

@SamSantala @Adobe @Photoshop Have you considered using photopea.com? It does not have 100% of the Photoshop functionality, sure. But it has the most important features. Give it a go. I switched years ago.

So am I reading this, right? @Adobe @Photoshop I can't use Photoshop unless I'm okay with you having full access to anything I create with it, INCLUDING NDA work?

"A random half of panelists were shown a CV and only a one-paragraph summary of the proposed research, while the other half were shown a CV and a full proposal. We find that withholding proposal texts from panelists did not detectibly impact rankings." link.springer.com/article/10.100…

@profpjm @vonKriegstein @Toffeemen68 I agree some would like to misuse it. However, we got a comment to compare our method with another class of methods (without proper citation). Then we conducted a lot of new experiments to satisfy this, only to get rejected because they had a slightly different paper in mind.

@vonKriegstein @Toffeemen68 you'll get/give proper guidance with a good action editor. reviewers telling authors what to specifically cite is not good

The peer review process.

Build what you need and use what you build. This is a core philosophy of my research. It shifts the focus away from publishing “papers” to what really matters — impact. This thread unpacks why I think this is a successful approach to science. 1/10 Or see: perceiving-systems.blog/en/post/build-…

@ngriffin_uk @Ax_Sharma So every time I open an issue and want to attach a screenshot, somebody would have to approve of that? I don't think this is a good idea. Why not just using random URL for anonymous attachments? No additional work for repo owners and no repo name in the link, hence no problem.

@Ax_Sharma I think this is actually quite a relatively simple fix, just require approvals / known contributors to upload as they do for status checks.

A GitHub flaw lets attackers upload executables that appear to be hosted on a company's official repo, such as Microsoft's—without the repo owner knowing anything about it. The following URLs, for example, make it seem like these ZIPs are present on Microsoft's source code repo: https://github[.]com/microsoft/vcpkg/files/14125503/Cheat.Lab.2.7.2.zip https://github[.]com/microsoft/STL/files/14432565/Cheater.Pro.1.6.0.zip But they are not. These ZIPs are #malware. An attacker, while commenting on any GitHub commit/PR, can "attach" a file that gets assigned a URL slug containing the name of the repo where the comment was made. Even if the comment is never actually posted or later deleted by the attacker, the link to the file remains live! And, the repo owner (Microsoft in this case) would have no knowledge of or control over such files. Threat actors have been abusing this flaw to distribute malicious executables under the false pretense that these are coming from credible organizations' code repos.

@loquaciousidiot @NC_Renic @scottjshapiro Well maybe the solution is to not write stuff like this at all. I learned never to write comments into the text itself. Even LaTeX comments are not good as this is visible at arXiv. 1. Do not write anything disrespectful, 2. If you must vent, do so over a coffee.

@NC_Renic @scottjshapiro I highlight the heck out of something like this when I write it so it doesn’t accidentally pop in

Editor: “please check the proofs carefully as this is the last opportunity to correct errors before publication” Academics: “yeah, looks sweet 🤘”

@phillip_isola Looks great! Is there a pdf version?

Our computer vision textbook is released! Foundations of Computer Vision with Antonio Torralba and Bill Freeman mitpress.mit.edu/9780262048972/… It’s been in the works for >10 years. Covers everything from linear filters and camera optics to diffusion models and radiance fields. 1/4

@chriswolfvision @JeromeRevaud I think this would make a great example next to the dictionary definition of the term "friendly academic shitposting".

@JeromeRevaud The funny thing is, in the paper he apologized to IBM and didn't want to show disrespect. Then, nukes their claim of quantum superiority by implementing a classical algorithm on a non-pipelined 8bit CPU w/o float instructions running at 1 Mhz, from the early 80ies.

SIGBOVIK never disappoints... More of this please!

@ylecun Signal.

WhatsApp vs iMessage is the new metric vs imperial. The entire world uses WhatsApp, save a few people who are iPhone-clutching Americans or from a country where WhatsApp is banned (like China). Yet iMessage users will actually claim it's objectively better, just like they will claim that Farenheit is intrinsically better than Celsius and inch better than cm, whereas it's just that they grew up with it.