- From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
- Date: Mon, 8 Jan 2024 16:10:58 -0500
- To: paoladimaio10@googlemail.com, W3C AIKR CG <public-aikr@w3.org>
The inferencing that goes on in Wikidata is, as far as I know, done by "bots" - external programs that look over information in Wikidata and add in some simple consequences of that information. The most active of these bots is KrBot, which uses information in Wikidata talk pages. See https://commons.wikimedia.org/wiki/File:DMD2023_-_A_better_way_to_enforce_a_data_model._Suggestions_to_improve_Autofix.pdf for a recent presentation on some of the problems with this methodology. I would very much like to see a better method for performing inference in Wikidata, or at least discovering violations of the constraints implied by the intended meaning of central Wikidata entities. As far as I know there is little direct use of Wikidata in old-fashioned AI. My current work with the Wikidata ontology is investigating and fixing problems with the building blocks of the ontology, including disjointness violations and violations of class orders. But I'm very interested in coming up with a good way of automatically performing inference and constraint checking in Wikidata. The Wikidata Ontology Cleaning Task Force is meeting Tuesday, 9 January 2024 at 1:30pm ET. https://calendar.google.com/calendar/u/0/embed?src=90c4d393de6b606fdb90ba6da1f7c1cc7afdd084df7776c02e55170620807843@group.calendar.google.com has information on how to join. peter On 1/6/24 05:49, Paola Di Maio wrote: > Greetings, Peter PS > > Although not strictly an AI topic, I share your email with @everyone @W3C AI > KR CG mailing list because > > 1. Its an opportunity to work with Peter who is an early proponent of > intelligent technologies and knows quite a great deal of stuff > > 2. Wikidata has been a good examples of plain SW implementations that offers > simple but robust (semantic) inference mechanism and is open source > Minimally addressed here > https://opendata.stackexchange.com/questions/15780/does-wikidata-offer-inferencing-reasoning <https://opendata.stackexchange.com/questions/15780/does-wikidata-offer-inferencing-reasoning> > > Peter, Please share with us any demo that shows wikidata reasoning/inferencing > capability? Whats the plan for leveraging wikidata to do AI? > > I take this opportunity to welcome all new members to this WG and encourage > you to introduce yourselves tot he list, share what you are working on, > how you plan to contribute etc > > Best greetings > > Paola DM, Chair > > > > On Thu, Jan 4, 2024 at 4:38 PM Peter F. Patel-Schneider > <pfpschneider@gmail.com <mailto:pfpschneider@gmail.com>> wrote: > > Wikidata (https://www.wikidata.org/wiki/Wikidata:Main_Page > <https://www.wikidata.org/wiki/Wikidata:Main_Page>) is a large (close > to 110 million entities) open-source repository of information. Wikidata uses > a data model that is similar to, but more general than that of either RDF or > labelled property graphs. Wikidata incorporates a large ontology using > regular > properties, as does RDFS. > > The Wikidata ontology has become very large, somewhere around 4 million > class-like entities. As for the rest of Wikidata, the Wikidata ontology is > the > result of edits from many different sources and by many different agents. > As a > result, there are many problems in the ontology. > > There have been several investigations and surveys about the Wikidata > ontology > and thus an increased awareness of the problems in the ontology. See > https://www.wikidata.org/wiki/Wikidata:Ontology_issues_prioritization > <https://www.wikidata.org/wiki/Wikidata:Ontology_issues_prioritization>, > https://www.wikidata.org/wiki/Wikidata_talk:Ontology_issues_prioritization#Overview_of_potential_solutions <https://www.wikidata.org/wiki/Wikidata_talk:Ontology_issues_prioritization#Overview_of_potential_solutions>, > and > https://commons.wikimedia.org/wiki/File:Wikidata_Challenges_in_Semantic_Web_Community.pdf <https://commons.wikimedia.org/wiki/File:Wikidata_Challenges_in_Semantic_Web_Community.pdf> > for more information. There is a task force starting up to address issues in > the Wikidata ontology. See > https://www.wikidata.org/wiki/Wikidata:WikiProject_Ontology/Cleaning_Task_Force <https://www.wikidata.org/wiki/Wikidata:WikiProject_Ontology/Cleaning_Task_Force> > for more information. > > If you are interested in analysis of the Wikidata ontology or in helping to > improve the ontology please contact me. Any kind of interest is welcome, > ranging from theoretical analyses of the ontology, to techniques for > reasoning > in Wikidata, to implementation of tools that help improve the ontology, to > direct editing of the ontology. I can assist you in finding out more about > the > ontology, introducing you to others involved with the ontology, forming > groups > that can address issues with the ontology, or developing a topic suitable for > academic investigation. > > Peter F. Patel-Schneider >
Received on Monday, 8 January 2024 21:11:05 UTC