Does it exist? (crowd sourced Wish List)

[zacharywhitley]

Is it not on the list? Do you wish that it was? Well add it here and myself or someone else might just go find it. If we can't find it than it's a good candidate for something to go and create.

Sometimes wishes really do come true!

[maximveksler]

I'm missing an interactive introduction to RDF going from triplet to Datalog that can act as killer demo for new people to grasp possibilities.

[maximveksler]

curriculum for graph theory, rdf, modelling. Real life example of walk throughs of modeling, semantic meaning, explained with lessons learned and co.

[zacharywhitley]

Interesting. Lets start another ancillary page on example-curricula.md and we can expand on the education/how-to/books section while we're at it. I wonder if there are any existing curricula we can use as a template.

• https://www.sti-innsbruck.at/teaching/curriculum/semantic-web
• https://www.phusewiki.org/wiki/index.php?title=Semantic_Technology_Curriculum
• https://www3.cs.stonybrook.edu/~pfodor/courses/CSE595/syllabus.html
• https://www.cs.unb.ca/~boley/cs6795swt/syllabus.html
• https://cs.hofstra.edu/~knarig/SemanticWeb/

[zacharywhitley]

"Does it exist" is a better title but I still wanted something to capture "I know it doesn't exist but wish it did" as an alternative to some of the stuff listed at EasierRDF, maybe "It doesn't exist....but I wish it did."

[maximveksler]

Updated the title.

n.b. https://github.com/w3c/EasierRDF seem like a wonderful resource. If I may ask - how did you hear about that?

[maximveksler]

n.b. II - a better format then a single issue might be a forum for discussion on a "per wish", could a forum on https://discuss.graph.vodka be valid option?

We could put a header on the awesome README.md list inviting people to participate and contribute.

[zacharywhitley]

EasierRDF seems to be a mixed bag. There seem to be some good intensions that are unfortunately mixed up with some misunderstandings. I'm sympathetic to the "lets make this easier to use" but don't like the "lets make this easier for me to use by getting rid of the parts I don't understand"

[pfodor]

These are the slides for the materials in my syllabus linked above: https://www3.cs.stonybrook.edu/~pfodor/courses/cse595.html If you need the powerpoint lectures, then I am happy to provide them. I keep the homework assignments in Blackboard as requested by my university, but I can send them by email if requested: pfodor at cs dot stonybrook dot edu

[rkranendonk]

I'm looking for an ontology that describes business processes. You already listed the W3C 'org' ontology but that's limited to the structure of the organization.

[rkranendonk]

Also looking for an ISO27001 / 27002 ontology.

[zacharywhitley]

It looks like your looking for more of a cyber ontology rather than a bussiness organization ontology (I can see w3c org being used in a iso27001 mapping.

I've actually started a separate awesome list specifically for ontologies at https://github.com/semantalytics/awesome-ontologies

It's just getting started to it needs some work. There are several papers by a Stephan Fenz on creating an ontology based on ISO27001. Unfortunately the link to the ontology is dead. It looks like they were running webprotege.

[zacharywhitley]

You might have some luck emailing him directly at stefan.fenz@tuwien.ac.at

[rkranendonk]

Thanks! The application area I'm looking into is in fact the GDPR. In my approach, the starting point is the processes that deal with personal data. These processes cause privacy risks, that need to be mitigated by "adequate measures", which usually have behavioural, governance and/or cybersecurity components. Compliance departments and certification agencies commonly handle GDPR as an add-on to ISO27001.

[zacharywhitley]

Cool. How about this? Ontology for EU's General Data Protection Regulation (GDPR)

https://www.w3.org/community/dpvcg/wiki/Data_Protection_Ontology_by_Bartolini_et._al https://www.linkedin.com/pulse/gdpr-ontological-primer-rémy-fannader https://www.w3.org/community/dpvcg/wiki/GDPRtEXT

[rkranendonk]

Thanks again Zachary. That Stefan Fenz is a real goldmine, Google-wise.

[zacharywhitley]

Yes, lots of papers on Stefan's work but unfortunately no ontology :(

From what I can find is that it was posted to sba-research.org with web-protege but since it was a web application and not simply the owl file it isn't available through web.archive.org. It seems as though Stefan wet on to found Xylem Technologies and became the basis for their risk and compliance management software AURUM

The work on this ontology was sponsored through various government grants so there might be some possibility that they are publicly available through as the deliverables of those grants but I've been unable to find anything.

Hope this helps. It's been fun trying to track this down and learning a little bit about ISO 27001. It seems like a good application of semantic web technologies.

[rkranendonk]

He went over to the dark side ;-)

[imrankhan1984]

Is there good xml/xsd to owl/rdf converter? Most industry standards like OPC UA information model and IEC61850 are available as xml/xsd files, and for non-domain experts it is very hard to create ontology models and ontology instances easily. I'm yet to find any suitable converter that can help non experts.

[m1ci]

@imrankhan1984

Is there good xml/xsd to owl/rdf converter? Most industry standards like OPC UA information model and IEC61850 are available as xml/xsd files, and for non-domain experts it is very hard to create ontology models and ontology instances easily. I'm yet to find any suitable converter that can help non experts.

Have you tried https://rml.io/docs/rml/tutorials/xml/?

[zacharywhitley]

If I understood @imrankhan1984 's question he's looking for a transform between the XML schema and OWL. There have been various efforts on this but I can't think of any that are current. This is the first one that comes to mind although I seem to remember others http://rhizomik.net/html/redefer/xsd2owl/

[imrankhan1984]

@m1ci and @zacharywhitley , Thanks for the link and comments. The issues that we face (and still there is no easy solution yet) concerns both XML and XSD files. I will check the rml link.

[m1ci]

@imrankhan1984 for XML you should defenitely try RML, it is under an active development. As for XSD, you could also try RML but not sure if you will get good results. Maybe a custom script would make most sense for the XSD.

[zacharywhitley]

@imrankhan1984 Try taking a look at ontmalizer. It's a bit old, last commit was a couple of years ago but might be worth looking at. There's a section in the awesome list for mapping data XML=>RDF.

[xkortex]

Are there any decent ontologies/schemas for machine learning dataset annotation? E.g. image labeling, bounding boxes, segmentation masks, etc?

[zacharywhitley]

That's a really good question that I've actually put a lot of thought into. The only current thing out there that I know if is image snippets which uses the lightweight image ontology but it's not really focused on ML. It doesn't help that ML uses semantic image annotation, but with a different meaning than semantic web, which make searching difficult. I think this would be a great application of semantic web technologies. You might also want to look into IIIF. It's a little confusing because it uses JSON-LD but very few of the implementations, if any, use the semantic part and just treat it as regular JSON. If you're using it for ML application there may be issues of scalability depending on the size of image corpus you're looking to work with. The w3c did some work on image annotation a long time ago that you can find here and the work on the web annotation data model and media fragments would be of use of you were to build your own system.

I think the killer application would be combining multiple annotated datasets. If I want pictures of cats and dogs I'm not too concerned about where they come from aside from data quality issues. How many face datasets are there? It would also save me a ton of time trying to understand your crazy ML data format. Is that box upper right/lower left? center/height/width? what coordinate system are you using?

I hope that helps.

[xkortex]

I think the killer application would be combining multiple annotated datasets. If I want pictures of cats and dogs I'm not too concerned about where they come from aside from data quality issues. How many face datasets are there?

This is literally what we deal with quite a lot. We've actually built off of MSCOCO Data Format, but since it's geared for Common Objects, it doesn't fit nicely for, well, uncommon objects. There are some implicit assumptions about the number of keypoints since it was designed around humans. Etc. We've also found that the shape of the schema (Dict[List] at the top level, e.g. {"images": [...], "annotations": [...]}) leads to performance issues, which would be alleviated by breaking into multiple files that link together. Stuff like that.

It would also save me a ton of time trying to understand your crazy ML data format. Is that box upper right/lower left? center/height/width? what coordinate system are you using?

I know this pain. So. Many. Bbox. Formats. Throw in other annotations types, and then geographic data...oof.

Good to know there is a demand for this sort of thing!

{scheme}://{server}{/prefix}/{identifier}/{region}/{size}/{rotation}/{quality}.{format}

This is actually a really cool spec, very heavy but I will definitely dig into this. Great find! A little strange they didn't use URI fragments for this though.

[zacharywhitley]

A little strange they didn't use URI fragments for this though.

There's actually a specific W3C recommendation for it too, W3C Media Fragments It's supported by Apache Marmotta but I'm not sure if anything has been done with Marmotta in a while.

If you think that's cool you might be interested in the rest based service api from ImgIX. It's not semantic web based but it's awesome.

[rtroncy]

There are many more implementations of the W3C Media Fragments URI specification, both server and client side (including native support in some web browsers and polyfill for others). The key use case of this recommendation is however the temporal dimension, the spatial one being limited to fixed rectangular bounding box in the frame.

There has also been proposals for extensions, such as those ideas (see github repo) from the MICO project and other ideas to bring dynamic media fragments.

One of the first semweb-by project around media fragments is probably Ninsuna. This slidedeck provides also useful pointers too. It has been followed by the LinkedTV project and the now running MeMAD project where I maintain an RDF Knowledge Graph making use heavily of the EBU Core ontology, Media Fragments and Web Annotations.

[zacharywhitley]

There are many implementations of the media fragments specification but not many with semantic web oriented projects. MICO uses Marmotta. Marmotta hasn't had a release in 6 years and the dynamic fragments and media fragments 2.0 are around 5 years old. All good stuff but I think it's important to be clear about what the current state of these projects and proposals are.

[rtroncy]

True. I have updated my previous post with more links

[zacharywhitley]

Sweet links, thanks. I'll add them to the list. PR's get priority.

[tmciver]

I'd love to see a section like "Semantic Web Applications" that listed applications (web, desktop, mobile) that use semantic web technologies. I've been looking for PIM type applications around calendaring, contacts, etc. but haven't had much luck finding any.

[danielgbaena]

Are there any public dataset like the one from: https://hci.stanford.edu/cstr/reports/2012-03.pdf ?

[zacharywhitley]

The closest I can think if would be https://github.com/ibm-aur-nlp/PubLayNet

[DrGilion]

Are there any dart implementations for reasoners like Jena in Java? I'm building a flutter app and want handle the reasoning on the dart side and not split the implementation for Android and iOS and their respective languages. I can't seem to find any libraries on pub or elsewhere :(

[tyan]

Why there is no dotNetRdf among SHACL implementations?

[gatemezing]

It looks like your looking for more of a cyber ontology rather than a bussiness organization ontology (I can see w3c org being used in a iso27001 mapping.

I've actually started a separate awesome list specifically for ontologies at https://github.com/semantalytics/awesome-ontologies

Great! how do you include ontologies in the list compared to other ontologies in vocabulary catalogs, like LOV https://lov.linkeddata.es/dataset/lov/