this vid has 94 views in a year. Now with the world of humans waking up to the benefits of this tech I bet this and other related videos will become the most valuable resource out there. Thank you for providing this information. I am not in the field nor do I have ANY experience, but this is what matters now so I'm here. Lets make the future!
I watched this one few months back and I wanted to rewatch but since then it is unlisted for some unknown reasons, and took me reaaaaally long to find it again. For the sake of accessibility could you please make it public again?
Ok. So Am I correct that probably the biggest difference (besides cardinality or union operatars) of OWL vs RDF(S) is that it can capture/imply relationship accross multiple nodes, even if it's not explicitely defined, while RDF(S) cannot?
when I classify a human a kind of person, or into a category of age or income, and then run a reasoner to prove that somebody is in that category, then I am doing a data validation.
Well delivered. What really strikes me is how a prescriptive knowledge cannot deal with reality. @13:00 just imagine the PM suddenly resigns and the project still goes on to completion. Same with rule-based/grammar and natural language: real language is broken.
The challenge is that if you make the edge "is author of", and later you want an entity for "Author" (a role), you've painted yourself into a corner that you can't now connect the Author role to the word "author" in the edge. You basically can't do CLASS -> EDGE -> EDGE. That might not be a big deal in all scenarios, depending on your use cases, but it's something worth considering.
Actually you can do "CLASS" - "EDGE" - EDGE. Infact, more interestingly, you can do "EDGE-...-EDGE" with the step in between staying anonymous. What is very interesting is that OWL can help you infer new EDGE's (e.g if we say "child-of" is the inverse of "parent-of"; then if we define one relationship ( ELROND - parent-of -> EOWYN), and we query for the other edge (EOWYN - child-of -> ?) some tuple stores that support Inferencing (e.g. Stardog) will return "ELROND" even though it does not exist in the data.
A very disappointing presentation. I cannot blame the presentor but those who filmed the presention. For a comedian the slides are not important but for a techincal presentation they are 80% of the values. #Poor
Hi Paul. Thank you for your honest feedback. It pointed to a (long overdue) oversight from our part that is now fixed! You can now access the Presentation Slides here: www.slideshare.net/ConnectedDataLondon/nowledge-architecture-combining-strategy-data-science-and-information-architecture-to-transform-data-to-knowledge-at-nasa
Some of us had a bad audio day - sorry about that. We hope you found the discussion interesting though. Hint: we also have a podcast, and that discussion is available as a podcast, too Sound quality may be better there due to post-processing: pod.co/the-connected-data-podcast
Since lots of modern applications (Power Apps, for example) want databases as their data source, have you all made a follow-up to this video that outlines how to convert the .owl ontology into a database or triple store? That would be REALLY valuable, because it'd let us ontologists leverage our content in a lot of different applications?
Folks you have to watch this if you were as blur as me when it came to RDF, OWL and the 3 Name alphabet police. A very comprehensive talk by an expert in the field. And even more amazing is that I just spent 6 days on an AML course. Thanks Tara.