Laurent Thomas @LauLauThom@qoto.org

Yet another AI reproducibility / reliability / generalizability crisis, this time in medicine

https://www.nature.com/articles/d41586-023-00023-2

You’ve already heard that Elsevier is evil. That they know this much about you — and that they are selling this data not just to our institutions but to god-knows-who-else — is cause for real, material protest. They may not have a crass loud mouthed billionaire spokesbaby but they’re so, so much worse. Open source, open access alternatives exist. Get them as far out of your life as possible. https://eiko-fried.com/welcome-to-hotel-elsevier-you-can-check-out-any-time-you-like-not/

"Juno: a Python-based graphical package for optical system design" — A complete library that includes a graphical user interface to design and visualise arbitrary optical elements, set up wave propagation simulations and visualise their results.

Paper: https://arxiv.org/abs/2212.12540
Code: https://github.com/DeMarcoLab/juno

#NewPaper #Python #Simulation

<<Find this useful? Please boost so that others can benefit too 🙂>>

I’m doing my new-to-me species dance! This little twisted diatom is Helicotheca tamesis and I found it in my Great Yarmouth sample. I’m very happy.

#marinebiology #marinebiologyuk #northnorfolkcoast #northsea #sea #ocean #biology #phytoplankton #plankton #marinetaxonomy #wildlife #nature #diatom #diatoms #microscopy #microorganism

Now seems like a good time to talk about how far-right extremists are using the Overton Window theory to force their ideas into the mainstream. The premise here is that today’s fringe ideas/beliefs can become tomorrow’s normal by shifting the boundaries of acceptable discourse. You do that by introducing people to more & more extreme ideas, so everything else seems reasonable in comparison.

eg, when “praising Hitler” becomes the comparison point, low-key bigotry suddenly doesn’t seem so bad. 1/

Toot! Toot!
Einen schönen zweiten Advent!

The best thing about this place is that there are no ads. I don’t know why this isn’t shouted about far more. I dread looking at my emails because there’s ads all through the damn thing. Ads everywhere on Twitter and Instagram. It’s a routine ugliness, a chore we’ve been told we must endure. But it’s not true.

Just published
@ScienceMagazine

A universal influenza mRNA vaccine effective against all 20 lineages of influenza A and B, protective in experimental models
https://www.science.org/doi/10.1126/science.abm0271

Masto on my mind 😊

If you are interested in image segmentation, this paper might be of interest:

https://arxiv.org/abs/2210.11006

There is code available here:
https://github.com/uncbiag/simpleclick

A word of warning — the segmentation appears to be interactive (as the title of the paper says 😛) and so I’m guessing might not work with any automated/scripted approaches you’re considering …

#DeepLearning #ComputerVision #Segmentation

💡Are you a fan of #openscience and #foss? Consider volunteering to review for @JOSS_TheOJ! Open reviews are conducted via @github issues, and I have often had authors say it's the most positive/constructive review process they've experienced!

👉 joss.theoj.org/reviewer-signup.html

Echoing a post from a colleague:

*Four* job opportunities in Pete Bankhead's research group @EdinUni_IGC
to work on #QuPath

https://forum.image.sc/t/four-positions-in-the-qupath-team-research-software-engineer-postdoc/73705

Focus on developing #opensource software & #ai tools for #bioimageanalysis - making complex analysis easier/better/possible

#javafx #deeplearning #digitalpathology #RSEng #ux

Which governance instruments can we employ to promote #OpenScience? Based on the pyramid of #CultureChange by @briannosek here are governance instruments that can encourage researchers to engage in two important practices of open science: documenting and sharing research #data. https://renebekkers.wordpress.com/2022/11/09/governance-instruments-to-promote-open-science/