#swib23 — Public Fediverse posts

@pietercolpaert @hvdsomp @nichtich At #swib23, Steven Folsom introduced it in the talk "Supporting sustainable lookup services", see https://swib.org/swib23/programme.html#supporting-sustainable-lookup-services for abstract, slides and recording. At this year's #swib25 there might be a talk about LDES which could be good opportunity to mention similarities and differences between both approaches. ;-)

@pietercolpaert @hvdsomp @nichtich At #swib23, Steven Folsom introduced it in the talk "Supporting sustainable lookup services", see https://swib.org/swib23/programme.html#supporting-sustainable-lookup-services for abstract, slides and recording. At this year's #swib25 there might be a talk about LDES which could be good opportunity to mention similarities and differences between both approaches. ;-)

@hauschke @SemAntiKast Hmm, now I'm not sure what discussion you're referring to.

We've mainly tried LLMs for metadata extraction, not classification as such. I gave a talk about it at #SWIB24 with Pierre Beauguitte from Norway (and also a lightning talk at #SWIB23 ).

Code for the experiments can be found here: https://github.com/NatLibFi/FinGreyLit

Most experiments involve fine-tuning smallish LLMs. Haven't tried reasoning models yet.

@hauschke @SemAntiKast Hmm, now I'm not sure what discussion you're referring to.

We've mainly tried LLMs for metadata extraction, not classification as such. I gave a talk about it at #SWIB24 with Pierre Beauguitte from Norway (and also a lightning talk at #SWIB23 ).

Code for the experiments can be found here: https://github.com/NatLibFi/FinGreyLit

Most experiments involve fine-tuning smallish LLMs. Haven't tried reasoning models yet.

@osma At #SWIB23 we discussed with @SemAntiKast about some tests you did with LLM based classification, if I remember correctly. Do I remember correctly? If yes, did you document your tests, and did you try different approaches (reasoning models etc.)?

@osma At #SWIB23 we discussed with @SemAntiKast about some tests you did with LLM based classification, if I remember correctly. Do I remember correctly? If yes, did you document your tests, and did you try different approaches (reasoning models etc.)?

Ich bin jetzt Moderator bei https://metadaten.community/

Auf der letzten #swib (#swib23 ) wurde in meiner Breakout Session der allgemeine Wunsch nach besserem einrichtungsübergreifendem Austausch zwischen #metadaten-Praktiker:innen im Alltag deutlich.

Das Forum bietet für den deutschsprachigen Raum dazu eine Möglichkeit. Alle sind eingeladen, sich dort auszutauschen. Auch vermeintlich basic Fragen sind gern gesehen.

Im Heft 6-2023 #b.i.t.online mein Bericht von der #SWIB23 – 15th Semantic Web in Libraries Conference

https://b-i-t-online.de/index.php

#bitonline

@flavoursofopen @otacke Insgesamt zeigt Sarven mit dokie.li sehr schön, was alles mit einem dezentralen Ansatz und Web-Standards möglich ist, inklusive Annotationen: https://dokie.li/#annotate Müssten wir halt mal ernsthaft angehen, sowas umzusetzen, auch für OER. Sarven hat bei der diesjährigen #swib23 die Keynote gehalten: https://csarven.ca/presentations/inherently-social-decentralised-and-for-everyone. Ich fand es sehr inspirierend aber gleichzeitig frustrierend, dass es ganz andere Dinge sind, an denen wir arbeiten (https://openbiblio.social/@acka47/111051231139459775) .

@flavoursofopen @otacke Insgesamt zeigt Sarven mit dokie.li sehr schön, was alles mit einem dezentralen Ansatz und Web-Standards möglich ist, inklusive Annotationen: https://dokie.li/#annotate Müssten wir halt mal ernsthaft angehen, sowas umzusetzen, auch für OER. Sarven hat bei der diesjährigen #swib23 die Keynote gehalten: https://csarven.ca/presentations/inherently-social-decentralised-and-for-everyone. Ich fand es sehr inspirierend aber gleichzeitig frustrierend, dass es ganz andere Dinge sind, an denen wir arbeiten (https://openbiblio.social/@acka47/111051231139459775) .

Das neue Heft - b.i.t.online 6-2023 - ist online. Mit Reportagen von der #FBM23 @frankfurterbuchmesse , von der #SWIB23 – 15th Semantic Web in Libraries Conference, von der E-Learning@media @steffenmeier , über DGI: 75 Jahre und länger Engagement für den professionellen Umgang mit Information und Wissen u.a.
https://b-i-t-online.de/index.php

Das neue Heft - b.i.t.online 6-2023 - ist online. Mit Reportagen von der #FBM23 @frankfurterbuchmesse , von der #SWIB23 – 15th Semantic Web in Libraries Conference, von der E-Learning@media @steffenmeier , über DGI: 75 Jahre und länger Engagement für den professionellen Umgang mit Information und Wissen u.a.
https://b-i-t-online.de/index.php

Das neue Heft - b.i.t.online 6-2023 - ist online. Mit Reportagen von der #FBM23 @frankfurterbuchmesse , von der #SWIB23 – 15th Semantic Web in Libraries Conference, von der E-Learning@media @steffenmeier , über DGI: 75 Jahre und länger Engagement für den professionellen Umgang mit Information und Wissen u.a.
https://b-i-t-online.de/index.php

Das neue Heft - b.i.t.online 6-2023 - ist online. Mit Reportagen von der #FBM23 @frankfurterbuchmesse , von der #SWIB23 – 15th Semantic Web in Libraries Conference, von der E-Learning@media @steffenmeier , über DGI: 75 Jahre und länger Engagement für den professionellen Umgang mit Information und Wissen u.a.
https://b-i-t-online.de/index.php

You can now find the links to slides and recordings from #swib23 on the programme page: https://swib.org/swib23/programme.html @swib

@margoline @elleneuler @RenkeSiems @zuphilip @hauschke @anjaoberlaender @dfg_public Eine wirklich offene Infrastruktur basiert auf offenen Standards: Protocols not platforms. Deshalb bestünde ein großer Teil der Implementierung in der Identifizierung bestehender Standards und ggf. Spezifikation neuer — für Metadaten/semantisches Markup, Reviews u.a. Annotationen, Archivierung. Es gibt da bereits vieles Vorarbeit, siehe etwa die Slides von Sarvens #swib23 Keynote: https://csarven.ca/presentations/inherently-social-decentralised-and-for-everyone

@margoline @elleneuler @RenkeSiems @zuphilip @hauschke @anjaoberlaender @dfg_public Eine wirklich offene Infrastruktur basiert auf offenen Standards: Protocols not platforms. Deshalb bestünde ein großer Teil der Implementierung in der Identifizierung bestehender Standards und ggf. Spezifikation neuer — für Metadaten/semantisches Markup, Reviews u.a. Annotationen, Archivierung. Es gibt da bereits vieles Vorarbeit, siehe etwa die Slides von Sarvens #swib23 Keynote: https://csarven.ca/presentations/inherently-social-decentralised-and-for-everyone

@hauschke Vor allem würde ich mir wünschen, dass das Web mal richtig genutzt und mit so einigen Paradigmen – zumindest erstmal experimentell – mal gebrochen wird: zum Beispiel mal was ohne Journals, PDFs, mit reichem semantischem Markup und standard-basierten Annotiationstools. dokie.li zeigt was möglich ist, & das gibt es seit zehn Jahren... Auch das Thema Discovery könnte mensch hier mal überdenken. (Siehe auch meine bittere Erkenntnis nach Sarven's #swib23 Keynote: https://openbiblio.social/@acka47/111051231139459775 )

@hauschke Vor allem würde ich mir wünschen, dass das Web mal richtig genutzt und mit so einigen Paradigmen – zumindest erstmal experimentell – mal gebrochen wird: zum Beispiel mal was ohne Journals, PDFs, mit reichem semantischem Markup und standard-basierten Annotiationstools. dokie.li zeigt was möglich ist, & das gibt es seit zehn Jahren... Auch das Thema Discovery könnte mensch hier mal überdenken. (Siehe auch meine bittere Erkenntnis nach Sarven's #swib23 Keynote: https://openbiblio.social/@acka47/111051231139459775 )

@swib presentations of #swib23 are now wikidatafied and looking for some P921 main subjects.

also some speakers have a need for their own #wikidata item.

swib23-scholia-page: https://scholia.toolforge.org/event/Q115590131#presentations

Recording of my #swib23 lightning talk "Extracting metadata from grey literature using large language models" now available: https://www.youtube.com/watch?v=wOzWG60SS_g

#metadata #llm #ai #cataloguing #llama2 #gpt

Recording of my #swib23 lightning talk "Extracting metadata from grey literature using large language models" now available: https://www.youtube.com/watch?v=wOzWG60SS_g

#metadata #llm #ai #cataloguing #llama2 #gpt

You can now find the links to slides and recordings from #swib23 on the programme page: https://swib.org/swib23/programme.html Enjoy.

Thanks all participants, speakers and workshop facilitators for another great SWIB conference. Bye everyone and travel save!

Next SWIB will happen 25-27 November 2024, with a core online event. We will keep you up to date on the developments regarding complementary decentral in-person events. #swib23 #swib24

Do you know more [Open Source] Knowledge Graph management systems such as Wikibase, Vitro, Web Protégé?

Only requirements are Web UI to edit both Ontology (TBox) and Statements (ABox) and RDF export. #swib23

Do you know more [Open Source] Knowledge Graph management systems such as Wikibase, Vitro, Web Protégé?

Only requirements are Web UI to edit both Ontology (TBox) and Statements (ABox) and RDF export. #swib23

Very clear and detailed presentation by @TinaTrillitzsch on how they switched their automated indexing from a legacy system to #Annif

#swib23 #ai #code4lib #metadata #indexing #skos

Very clear and detailed presentation by @TinaTrillitzsch on how they switched their automated indexing from a legacy system to #Annif

#swib23 #ai #code4lib #metadata #indexing #skos

The "general-purpose semantic web application Vitro" presented at #swib23 (I'd call it a knowledge graph management system) may be an alternative to Wikibase for some LOD projects. Unfortunately documentation is limited: https://github.com/vivo-project/Vitro

The "general-purpose semantic web application Vitro" presented at #swib23 (I'd call it a knowledge graph management system) may be an alternative to Wikibase for some LOD projects. Unfortunately documentation is limited: https://github.com/vivo-project/Vitro

Kai Labusch from @stabi_berlin is presenting at #swib23 a very sophisticated service for named entity recognition and entity linking in digitized texts, constructed with machine-learning based on #Wikidata entities and their mentions in #Wikipedia. I wonder if one could/should use the #Reconciliation Service API protocol for use cases like this one. I guess the protocol would have to allow to send the context sentence along with the NERed string we want to reconcile. /cc @pintoch

Kai Labusch from @stabi_berlin is presenting at #swib23 a very sophisticated service for named entity recognition and entity linking in digitized texts, constructed with machine-learning based on #Wikidata entities and their mentions in #Wikipedia. I wonder if one could/should use the #Reconciliation Service API protocol for use cases like this one. I guess the protocol would have to allow to send the context sentence along with the NERed string we want to reconcile. /cc @pintoch

Linked Open Data workflow for #Wikidata as just mentioned by Will Kent from WikiEdu at #swib23: https://www.wikidata.org/wiki/Wikidata:Linked_open_data_workflow

Linked Open Data workflow for #Wikidata as just mentioned by Will Kent from WikiEdu at #swib23: https://www.wikidata.org/wiki/Wikidata:Linked_open_data_workflow

At #swib23 yesterday heard about connecting archival metadata with bibliographic data formats. Today connecting archival metadata with Wikidata data model: https://www.wikidata.org/wiki/Wikidata:WikiProject_University_of_Washington_Archival_Metadata

At #swib23 yesterday heard about connecting archival metadata with bibliographic data formats. Today connecting archival metadata with Wikidata data model: https://www.wikidata.org/wiki/Wikidata:WikiProject_University_of_Washington_Archival_Metadata

Accompanying my colleagues' presentation "Entity linking historical document OCR by combining Wikidata and Wikipedia" on #SWIB23 today, we published our three Named Entity Disambiguation & Linking models on Hugging Face (de/en/fr) at https://huggingface.co/SBB as well as three accompanying training databases on Zenodo:

https://doi.org/10.5281/zenodo.7767403 (de)

https://doi.org/10.5281/zenodo.7773986 (en)

https://doi.org/10.5281/zenodo.7773745 (fr)

@cneud @stabi_berlin

#ner #namedentitylinking #namedentitydisambiguation

Accompanying my colleagues' presentation "Entity linking historical document OCR by combining Wikidata and Wikipedia" on #SWIB23 today, we published our three Named Entity Disambiguation & Linking models on Hugging Face (de/en/fr) at https://huggingface.co/SBB as well as three accompanying training databases on Zenodo:

https://doi.org/10.5281/zenodo.7767403 (de)

https://doi.org/10.5281/zenodo.7773986 (en)

https://doi.org/10.5281/zenodo.7773745 (fr)

@cneud @stabi_berlin

#ner #namedentitylinking #namedentitydisambiguation

Interesting distributed terminology service Network of Terms, found in the #swib23 feed, Thanks to @ennomeijers Code and documentation available on #GitHub #terminology #semanticweb #graphql #nfdi https://github.com/netwerk-digitaal-erfgoed/network-of-terms-tutorial/wiki/Goal-and-purpose-of-the-Network-of-Terms

I had to leave #swib23 early, but I had a really good time! I even held a little Lightning Talk about integrating Network of Terms into Cocoda Mapping Tool. Slides here: https://coli-conc.gbv.de/publications/2023-09-12-SWIB23-Lightning-Talk.pdf

Thanks everyone! 👋

Read more about the #SKOS reconciliation service and a first version of a SKOS Shacl shape in the slides to the #swib23 lightning talk @acka47 just gave: https://pad.gwdg.de/p/0ytHDo597

My #swib23 lightning talk "Extracting metadata from grey literature using large language models"
Slides: https://tinyurl.com/swib23-metadata-extraction

#metadata #llm #llms #ai #gpt #llama2 #cataloguing

Come to the Dietrich-Bonnhoefer-Saal if you have interest in #SkoHub . @acka47 will be around for ~30 minutes. #swib23

@osma I've added #SkoHub, #Metafacture and #lobid stickers at the registration desk. Feel free to get some. #swib23

@osma I've added #SkoHub, #Metafacture and #lobid stickers at the registration desk. Feel free to get some. #swib23

@osma I've added #SkoHub, #Metafacture and #lobid stickers at the registration desk. Feel free to get some. #swib23

@osma I've added #SkoHub, #Metafacture and #lobid stickers at the registration desk. Feel free to get some. #swib23

@osma I've added #SkoHub, #Metafacture and #lobid stickers at the registration desk. Feel free to get some. #swib23

TIL from the #swib23 talk "Wikibase as an institutional repository authority file" given by Michael Lindsey that there is a Wikibase cloud hoster in the making: https://www.wikibase.cloud/

We recently have been developing a generic #SKOS #reconciliation service (still in early beta). You upload a turtle file of a SKOS vocab and get back the URL of a reconciliation endpoint. This might also come handy in different use cases presented at #swib23. @acka47 will try to give a short demo later.