mastouille.fr est l'un des nombreux serveurs Mastodon indépendants que vous pouvez utiliser pour participer au fédiverse.
Mastouille est une instance Mastodon durable, ouverte, et hébergée en France.

Administré par :

Statistiques du serveur :

592
comptes actifs

#replicability

0 message0 participant0 message aujourd’hui
Suite du fil

TODAY (Mon 16 June) at 16:00 CEST, we are delighted to welcome not one, not two, but three guest speakers to #ReproducibiliTea in the HumaniTeas: @xenia_ks and Anna Yi Leung (both from the Self Learning Systems Lab of the University of Cologne) and @JohannesBreuer (from GESIS—Leibniz Institute for the Social Sciences, Cologne)! ☀️

They will lead a session on the role of language and its role for #replicability based on their recently published short paper (doi.org/10.1057/s41599-025-043). 📖

As usual, you can join us in Cologne @unibibkoeln for a cuppa or via Zoom (join our mailing list to get the link: lists.uni-koeln.de/mailman/lis). 🫖 #OpenScience #academia #linguistics @johannesbreuer.com

Les interventions de la journée d'étude ARDoISE du 17 décembre sont en ligne et accessibles depuis le site ScienceConf : je2024-ardoise.sciencesconf.or #reproductibilté #replicability #ResearchSoftware Interventions de @zimoun @rougier @NaudetFlorian , Isabelle Blanc administratrice ministérielle des données, code source et logiciels, et d'autres encore...

je2024-ardoise.sciencesconf.orgJournée d'étude Ardoise - 17 décembre 2024 - Sciencesconf.org

⚙️ Another significant step toward embracing #OpenScience in Ukraine! 🇺🇦

⌛ The Ukrainian Reproducibility Network #UARN is launching during the Ukrainian Open Science Forum on November 21-23, 2024 in Lviv!

As part of the UARN, Local Networks will be established in different regions of Ukraine to support & promote #transparency #reproducibility #replicability practices across all disciplines.

⌛ There’s still a chance to join either in person or online: peers.international/uarn?fbcli
#UOSF2024

The state of scientific #replicability in AI for large models is quite depressing. It seems to be currently well-accepted that, given that very few actors in the world can afford to replicate model training (which is certainly true), then it is OK to not release training code for your experiments to reviewers and the broader scientific community. I vehemently disagree and, frankly, consider this a mire in how much we can trust associated "scientific" results.

Last week I attended the 6th Perspectives on Scientific Error Conference at @TUEindhoven
I learned so much! About #metascience #preregistration #replicability #qrp questionable research practices, methods to detect data fabrication, #peerreview, #poweranalysis artefacts in #ML machine learning...
I'm impressed by the commitment of participants to improve science through error detection & prevention. Thanks to the organizers Noah van Dongen, @lakens @annescheel Felipe Romero and @annaveer

:blobcatread: A Survey on the Attitudes Towards and Perception of Reproducibility and Replicability in #Sports and #Exercise Science

"Statistical education should be prioritised for early career researchers which could positively affect publication and peer review. Researchers must accept responsibility for #reproducibility and #replicability with thorough project design, appropriate planning of analyses, and transparent reporting practices"

storkjournals.org/index.php/ci