#recommender — Public Fediverse posts

The research on very large online platforms (VLOP) is today unnecessarily complicated. #CoCoDa, a joint initiative by Université de Lausanne Maastricht University Open Data Institute and Universität St.Gallen (HSG) collects very useful tools and cutting-edge research to shed light on the work of #recommender systems and audit those platfoms.

Well done Aurelia Tamo-Larrieux, Simon Meyer et al!

https://snsf-cocoda.github.io/

🎁 BENEFITS to RESEARCHERS of being a PCI RR #recommender cont.

8) As a PCI RR recommender, we follow PCI RR policies on selecting submissions based on scientific validity (rather than subjective impact) and addressing implicit biases. This facilitates making academia more fair and improves the quality and rigor of published research

9) We receive training in how to be a recommender (editor) and how to handle registered reports

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

🎁 BENEFITS to RESEARCHERS of being a PCI RR #recommender cont.

7) As a PCI RR recommender, researchers oversee peer review of registered reports both within & beyond a particular journal, thus increasing global reach - constructively shape the research record at a supra-journal level

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

🎁 BENEFITS to RESEARCHERS of being a PCI RR #recommender cont.

5) As a PCI RR recommender, we contribute to an ethical publishing pathway and are NOT donating our time to a journal/publisher who exploits research and researchers and drains money out of academia

6) As a PCI RR recommender, we see a variety of registered reports. This experience helps us learn more about the process 🎓 and improves our own research and registered reports

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

🎁 BENEFITS to RESEARCHERS of being a PCI RR #recommender cont.

3) A PCI RR recommender’s recommendation counts as a research output. It's posted @ PCI RR website, has a DOI, is citable, & shows up in your google scholar profile

4) Recommenders join the PCI RR community to be part of something bigger 🌍 - making a new research sharing pathway that lets researchers take back control of the publishing process by pooling expertise and clout - independent of publishers

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

🎁 BENEFITS to RESEARCHERS of being a PCI RR #recommender cont.

2) PCI RR’s published review histories encourage helpful interactions between recommenders, authors, and reviewers. Recommenders feel they can help authors improve research & are excited to see the findings. It’s like the recommender & reviewers become stakeholders 🎉 in the research, rather than gatekeepers 🔑

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

🎁 What are the BENEFITS to RESEARCHERS of being a PCI RR #recommender?
(Thanks to Naho Konoike for sparking this thread! 🙏 )

1) PCI RR recommenders find it rewarding to handle RRs because their input can be incorporated into the research and make it publishable. This prevents having to reject non-RR articles because the methods couldn’t answer the questions - a waste of researcher time and public money

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

After enquiry about becoming a #recommender at PCI RR, and if you meet our criteria, we'll add you to our waitlist, which the Managing Board approves. Recommenders will be invited in waves balanced for diversity

⏩ Want to help speed up the process of getting off of the waiting list? Find colleagues from groups who are traditionally underrepresented & encourage them to email us their interest!

👉 What does a PCI RR recommender do? Check out https://spore.social/@pcirr/114850930291005956

ℹ️ https://rr.peercommunityin.org/help/become_a_recommenders

@phpmacher #RSS nutzen: Sicher. Immer noch populär. Ob als Eingangsformat für andere Medien oder direkte Nutzung durch Menschen.
Gibt diverse Systeme die lokal aggregieren und filtern, auch self-hostable. Teilweise über die eingebetteten #Metadaten, teilweise mit #Regex gegen andere Inhaltsfelder.
Der Schritt von da zu einer Art #Bayesfilter oder den NN wie sie #rspamd zur Klassifizierung nutzt oder sogar einem SLM/kleineren #LLM auf #Ollama, ggf. mit #TPU / #APU Support sollte überschaubar groß sein.
Eins der Probleme wird aber die zunehmend schlechte Feedqualität was Tags/Metadaten und der notorische (wenn gleich sehr verständliche) Hang zur Nicht-Auslieferung des Volltext im Feed sein. Da müsste man ggf auf Verdacht das Original fetchen, ggf. als zweite Stufe.

Und natürlich ist ein #Recommender-System was deutlich anderes als ein mehr oder minder ausgefeilter #Classifier. Aber lassen wir das...

What does a PCI RR #recommender of #RegisteredReports do?
1) Authors submit article & suggest several recommenders in their area & with whom they have no conflicts of interest. Once the authors click "Submit", invitations are automatically sent to suggested recommenders

2) Invited recommenders receive an email invitation to handle the submission. They can accept, decline, or ignore. Many people are invited at once, so invitees don't need to feel pressured to accept every invitation
1/n

Excited to announce that our article “Narrative co-creation with SPORE: a spatial #hypertext based #narrative #recommender system” has been published in the NRHM journal as #OpenAccess! 🎉 Co-authored with Daniel Roßner and Sam Brooker. Check it out here: https://doi.org/10.1080/13614568.2024.2402704
#elit #eliterature #recsys

Great idea amazon, I'd love some tomato puree with my cola! #recommender #AI #Amazon

Here is just a little #UX hint for the good people who develop #YouTube: I shouldn't be able to watch, literally, ONE fucking video that is unusual for me, and then have my entire recommendation feed completely changed.

Completely.

How can your system be that sensitive, and that delicate?

One video shouldn't change everything in my feed!

#uxdesign #recommendations #recommender #userexperience #uiux #uidesign #uiuxdesign

Manchmal sucht man ja nach #Katalogen, die einen #Recommender-Service bieten. Und wie immer fallen die einem nicht gleich ein. #bibtip hat eine Seite mit #Referenzen. https://www.bibtip.de/de/references.html

Inherent Limits on Topology-Based Link Prediction

Justus Isaiah Hibshman, Tim Weninger

Action editor: Laurent Massoulié.

https://openreview.net/forum?id=izL3B8dPx1

#graphs #recommender #graph

The latest Leviathan newsletter looks at the wider implications of the case against Google. Will it help or lead to consolidation of the open web?

There are more theories, this time by @henryfarrell, that it's us and not only algorithms that are causing misinformation & polarisation;

And the latest research on the impact on clinical assessment by AI.

And #privacy: It's complicated.

https://leviathan.substack.com/p/the-wider-implications-of-the-case

#techpolicy #recommender #Google #openweb #advertising #antitrust #misinformation

The Graph Cut Kernel for Ranked Data

Michelangelo Conserva, Marc Peter Deisenroth, K S Sesh Kumar

https://openreview.net/forum?id=SEUGkraMPi

#ranking #rankings #recommender

heute im dienstkalender [zumindest "mit einem ohr"]: der workshop zum abschluss des B!SON-projekts @oa_bison: https://projects.tib.eu/bison/en/news/detail/save-the-date-3-workshop-zum-projektabschluss-am-2512023/ #openaccess #recommender #service
RT https://openbiblio.social/@oa_bison/109748816510785490

#FollowGraph is a great example of why #Mastodon does not need #algorithms built in.

It is a stand-alone #openSource algorithm that is similar to the #follow #recommender on Twitter, except it is not pushed in your face, the algorithm is explicit and documented, it is #privacy preserving, and other developers can easily come up with their own variation for you to try.

And it's is a great way to find people to follow,

This is the benefit of an open platform.

https://followgraph.vercel.app/

Does #mastodon have any #recommender system plug-ins yet? For those who would consider opting in. Is a recommender system even possible without some kind of centralised data collection?

Ugh! “Related Tweets” is anything but. How do I turn off this useless & irritating feature in #Twitter? 😠

#WTF #UX #Recommender #RelatedTweets