Re: Reasoning with ontologies and knowledge graphs?

Hello,

We are using inferencing over SKOS and wikidata subclasses within 
searches for image data in ImageSnippets.

We build inference graphs when we save the triples used to describe images.

We also have a search paths function that outlines the hops taken across 
the datasets (DBpedia - SKOS broader/SameAs with Wikidata/subclasses as 
well as Art and Architecture thesaurus).

I don't know if this is what you are really asking for here, but if 
anything listed here, perhaps describes an informal approach that mimics 
human reasoning.

Our search paths function is also quite revealing about oddities that 
come up in places like the subclasses used in Wikidata for example - 
when you can do things like get a result of an image of a Bay in New 
Zealand for a search for a term like: 'communication medium'

The hops returned are as follows:

http://dbpedia.org/resource/New_Zealand (sameAs)
Wikidata: New Zealand http://www.wikidata.org/entity/Q664 a Commonwealth 
realm http://www.wikidata.org/entity/Q202686
subclass of kingdom http://www.wikidata.org/entity/Q417175 subclass of 
monarchy http://www.wikidata.org/entity/Q7269
subclass of monarchic system http://www.wikidata.org/entity/Q22676587 
subclass of form of government http://www.wikidata.org/entity/Q1307214
subclass of administrative type http://www.wikidata.org/entity/Q2752458 
subclass of classification system http://www.wikidata.org/entity/Q5962346
subclass of knowledge organization system 
http://www.wikidata.org/entity/Q6423319 subclass of communication medium 
http://www.wikidata.org/entity/Q340169

Another interesting result was finding images of native americans 
returned for a search on 'insects' --- (because both enties are related 
to 'tribes'. )

Best,

Margaret Warren






On 12/12/21 9:11 AM, Melvin Carvalho wrote:
>
>
> On Sat, 11 Dec 2021 at 12:00, Dave Raggett <dsr@w3.org> wrote:
>
>     I am a member of the AIOTI WG Standardisation activity on semantic
>     interoperability [1]. We are interested in getting a better
>     feeling for what kinds of automated reasoning people are
>     realising, or seeking to realise, with ontologies and knowledge
>     graphs, along with the associated use cases.
>
>     In principle, there are many opportunities for a wide variety of
>     different forms of reasoning, including logical deduction and
>     ontological entailment, induction, abduction, spatial and temporal
>     reasoning, causal reasoning, plausible reasoning, qualitative
>     reasoning, fuzzy reasoning, analogical reasoning and so forth.
>     This spans approaches based on formal semantics, approaches based
>     on probability theory, as well as informal approaches that mimic
>     human reasoning.
>
>     We welcome suggestions for designing a survey on automated reasoning.
>
>
> Thanks for posing this question,  It's something I've asked myself a 
> few times
>
> I've been using the semantic web for about 15 years for practical 
> projects, quite a lot for personal use
>
> From my experience, it's rare that I have seen inferencing is used 
> either by myself, or in projects I've come across, however here was 
> one tweet that I collected:
>
> https://twitter.com/bobdc/status/1318164165584891905
>
> It uses a python script to transform data, and give back certain 
> inferences
>
> We used smushing a bit and (Inverse) Functional Properties in the 
> tabulator project w/ rdflib.js, but not all that much
>
> I've seen domain and range discussed a few times, including in the 
> fediverse.  Perhaps this could used in conjunction with validations or 
> shapes.
>
> I think today the de facto use of the semantic web is schema.org 
> <http://schema.org> and iirc https://schema.org/Person for example, I 
> think no longer gives back machine readable data, only html (I could 
> be wrong)
>
> So that leads me to think the value of the semantic web, is in the 
> namespacing rather than inferencing, or even shared schemas
>
> A comment I heard on this recently: "make RDFS so light it's useless, 
> and OWL so complicated nobody uses it - always needed something in the 
> middle, that was clear from the get go", slightly tongue-in-cheek, but 
> that did resonate with me
>
> I think inferencing would benefit from a rethink, if it is to gain 
> more traction
>
> Very interested in practical experiences of others wrt inferencing, 
> particularly anything that has gained public traction
>
> Personally, Id like to see a future of inferencing to look a bit more 
> like the python script above, and able to do data transformations from 
> one form to another.  I think that could have lots of practical 
> applications
>
>
>     [1] https://aioti.eu/wg_standardisation/
>
>     Many thanks,
>
>     Dave Raggett <dsr@w3.org> http://www.w3.org/People/Raggett
>     W3C Data Activity Lead & W3C champion for the Web of things
>
>
>
>