elias.fediscience.org.ap.brid.gy
elias
@elias.fediscience.org.ap.brid.gy
1 followers 0 following 13 posts
research engineer at @L3S_Research_Center working on recommendations, NLP, bibliometrics, etc. [bridged from https://fediscience.org/@elias on the fediverse by https://fed.brid.gy/ ]
fediscience.org
Posts Media Videos Starter Packs
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · 12d
It seems like there's a growing number of taxonomies for how to disclose the use of generative AI in scientific writing. Here are the ones from STM, CEUR-WS and Leiden […]

[Original post on fediscience.org]
STM taxonomy with 9 different classes including description, examples and excluded activities CEUR-WS taxonomy with 13 classes and descriptions Leiden taxonomy with 8 macro level classes and 40 micro level classes
2 1
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · 15d
OA.Works is shutting down Open Access Button and InstantILL in November.
They recommend switching to Unpaywall as a replacement for OpenAccessButton.

https://blog.oa.works/sunsetting-the-open-access-button-instantill/
#Scientometrics
Sunsetting the Open Access Button & InstantILL
How we’re refocusing our work to serve the Open Access movement.
blog.oa.works
2
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · 16d
"OpenAlex Analytics alpha is paused." https://openalex.org/analytics

@OpenAlex has paused their Analytics Dashboard alpha because it's too expensive and they want to focus on the backend rewrite.
Hopefully, they'll continue their work in the future! I really […]

[Original post on fediscience.org]
Screenshot of Openalex' statement. The most important part: We’re pausing the alpha, effective now; you’ll no longer be able to use or query the Analytics. We’re doing this for two reasons: We’ve learned how expensive it is to answer these complex ad-hoc queries. Our current numbers suggest we’d need to charge you and other users about $50k/yr, and our research indicates there’s no market at this price point (that said, if we’re wrong and you’ve got $50k/yr for this, let us know…our minds can be changed 😁). I think we need to figure out a way to do Analytics more cheaply to make it viable—either use a cheaper technology or support less flexibility. We’ve got some ideas on both counts, but we need time to look into it. Which brings us to… We need to focus on the new OpenAlex rewrite launch over the next few months. This is the biggest project we’ve ever done and we need to really zone in or it’s not going to work. So we’re saying “no” to everything we possibly can in Q4, so that we can say a great big “yes” to making OpenAlex faster, more accurate, and more comprehensive (we’re passing 400M works!). We’ll return to Analytics in 2026 with a new approach, probably one integrated more tightly with the existing UI and definitely one taking advantage of our turbocharged new backend architecture.
1 8 1
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Sep 7
Just arrived in Granada for #esss2025.
Excited to learn more about #Scientometrics the next few days.
https://esss.info/
european summer school for scientometrics – esss.info
esss.info
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Aug 11
We've improved @oa_bison - the open-access journal recommender! Based on last year's research, we've updated our recommendation method:
- Switched from a custom BiLSTM to Qwen3-Embeddings
- Changed the Elasticsearch scoring
=> the recommendations should be more accurate and less biased by […]
Original post on fediscience.org
fediscience.org
1
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Aug 8
@cutterkom @OpenAlex if you mean the authors of all works in the "referenced_works", then you could use the "cited_by" filter. You would get the works which includes the "authorships" field.

For example like this […]
Original post on fediscience.org
fediscience.org
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Aug 8
@cutterkom @OpenAlex do you mean the "referenced_works" API field?
1
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Jul 11
We are currently working on a project that involves data annotation and received a few questions about the tools we are using. For those interested:
We use Label Studio, an open-source annotation tool that’s easy to set up:
$ pip install label-studio
$ […]

[Original post on fediscience.org]
Screenshot of the different annotation templates in label studio. On the left are the categories: Computer Vision, Natural Language Processing, Audio/Speech Processing, Conversational AI, Ranking & Scoring, Structured Data Parsing, Time Series Analysis, Videos and Generate AI. Natural Language Processing is selected and the following templates are shown: Question Answering, Text Classification, Named Entity Recognition, Taxonomy, Relation Extraction, Text Summarization, Machine Translation.
1
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Jun 30
Random discovery: The robots.txt file on SpringerLink explicitly disallows bots from accessing one specific article: “The Hypersexualization and Undesirability of Black/African Women.”
https://link.springer.com/chapter/10.1007/978-94-6209-446-8_8
The first part of the robots.txt with the values: User-agent: * Disallow: / Disallow: /chapter/10.1007/978-94-6209-446-8_8 Disallow: /content/pdf/10.1007/978-94-6209-446-8_8.pdf Allow: /$ Allow: /?$ Allow: /brands/ Allow: /journal* Allow: /subjects/* Allow: /subject/* Allow: /journal/*/collections* Allow: /collections/* Allow: /article$
3
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Mar 27
New tool by AllenAI: Paper Finder https://paperfinder.allen.ai

It's an LLM-powered literature search system based on their SemanticScholar data set and semantic search.
Ai2 PaperFinder
paperfinder.allen.ai
1
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Mar 23
@NickEast_IndieWriter in Braunschweig (Germany), there is a bookshelf that acts as a door to a bar if you pull on the right book.

https://www.ardmediathek.de/video/hallo-niedersachsen/besuch-in-der-geheimen-bar-in-braunschweig/ndr/Y3JpZDovL25kci5kZS83MjQzYWIzZC03ODdmLTQzZmMtODliZC03ZjNhZDY1MzMyNzQ
Hallo Niedersachsen: Besuch in der geheimen Bar in Braunschweig - hier anschauen
Nur wer am richtigen Buch zieht, kommt hinein ins Hide & Seek. Seit drei Jahren gibt es die Bar von Besitzer Michael Grazé.
www.ardmediathek.de
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Mar 5
@jonny did you try to reduce the jvm size? I do that pretty much every time I set up elasticsearch. Otherwise it takes up way too much memory.

https://www.elastic.co/guide/en/elasticsearch/reference/current/advanced-configuration.html
Set JVM options | Elasticsearch Guide [8.17] | Elastic
www.elastic.co
elias.fediscience.org.ap.brid.gy
elias @elias.fediscience.org.ap.brid.gy · Feb 25
Apparently aideadlin.es is no longer maintained according to this GitHub issue: https://github.com/paperswithcode/ai-deadlines/issues/723.

Instead, they point to ccfddl.com (GitHub repo: https://github.com/ccfddl/ccf-deadlines), which tracks "deadlines of conferences recommended by China […]
Original post on fediscience.org
fediscience.org
1