I'm very skeptical about the capability of computers to really understand natural language and extract knowledge and reasoning about this data. I think the AI field will have to improve a lot to have succesful results on reasoning in the world of semantic web.
I don´t really see the point in this. I work as a webdesigner. What´s the benefit for my customers if I write more code (= taking longer = costing more). Do the search engines like this kind of pseudo-semantic informations? Will contents of webpages be found easier and faster? I mean I like semantic html, but this... Does anybody understand what I mean?
This is just a way of creating a web of relations between object. You can define your objects and your relations. And this web can be queried very precisely like a database. Computational linguistics is trying to make tools to automatically create rdf out of content. And the main benefit that this is trying to bring to the web as you know it is to have the content of the web pages described in this manner. So imagine having such an rdf for all the content on the net.
And this huge web would become like a global brain. There already are solutions to translating a natural language question into a database style query. So the user will be able to naturally ask the computer complex questions and get precise responses. This google thing is just an example but the power of such technology is hard to describe. Governments could parse emails and phone calls. Having all this knowledge, AI will flourish because the main problem of AI is the lack of knowledge base.
glad the guy ended up writing "dave likes cookies" - until the dude wrote the 2nd o i thought he was gonna write something else & was taking the piss outta me
I do think though that it does add an extra layer of sophistication in terms of data manipulation, but I think it's still al long way off being truly semantic from the machine point of view.
So the machine knows Jane is a freid of Mac - does it know what Jane is? A girl? What's a girl - see where this is going?
Thanks for the foaf:name correction - I'll make sure to make the change it in future videos.
As for the mixing semantics/syntax comment, I think practicality plays a large role in RDFa. There is benefit for keeping the machine readable and the human readable data in sync and in the same document, which is what both Microformats and RDFa does.
But in your example, the human-readable text "Jane is friends with Mac" is not in sync with the machine-readable data "#jane foaf:knows #mac". You can change the text and the data could stay. Do you call this "in sync"? Aren't you asking people to keep text and data "in sync"? Is there a way to generate the text "Jane is friends with Mac" from the data "#jane foaf:knows #mac". It's not obvious from your example.
I also don't think that having more triples will help computers to "understand" anything. More triples will only allow computers to do simple computations on data, but not more.
This could be used by AI in the future, but by itself triples bring no understanding, IMO.
The point I was trying to make is that we have to lay the foundation first, which is exposing triples on web pages. Once those triples are there, there are many, many types of applications we can build based on that foundation. It is analogous to saying that the more people and pages there are on Wikipedia, the more useful that site becomes...
The next production is probably going to be an advanced RDFa tutorial. Also wanted to do one on the Microformats approach. We'll be doing several on Bitmunk as well...
If you've got any ideas/suggestions of what you'd like to see next, please let me know!
Super cool stuff! Thanks for sharing :)
carpediematlas 2 months ago
What a great Video .You simply Rock.............Keep the gud work Up
ifithegr8 1 year ago
I'm very skeptical about the capability of computers to really understand natural language and extract knowledge and reasoning about this data. I think the AI field will have to improve a lot to have succesful results on reasoning in the world of semantic web.
McAko 1 year ago
I don´t really see the point in this. I work as a webdesigner. What´s the benefit for my customers if I write more code (= taking longer = costing more). Do the search engines like this kind of pseudo-semantic informations? Will contents of webpages be found easier and faster? I mean I like semantic html, but this... Does anybody understand what I mean?
Platinger 2 years ago
This is just a way of creating a web of relations between object. You can define your objects and your relations. And this web can be queried very precisely like a database. Computational linguistics is trying to make tools to automatically create rdf out of content. And the main benefit that this is trying to bring to the web as you know it is to have the content of the web pages described in this manner. So imagine having such an rdf for all the content on the net.
radsimu 1 year ago
And this huge web would become like a global brain. There already are solutions to translating a natural language question into a database style query. So the user will be able to naturally ask the computer complex questions and get precise responses. This google thing is just an example but the power of such technology is hard to describe. Governments could parse emails and phone calls. Having all this knowledge, AI will flourish because the main problem of AI is the lack of knowledge base.
radsimu 1 year ago
glad the guy ended up writing "dave likes cookies" - until the dude wrote the 2nd o i thought he was gonna write something else & was taking the piss outta me
(must be paranoid cos i don't have a girlfriend)
lancsFrogger 2 years ago
Good overview of RDFa and why it's important.
I do think though that it does add an extra layer of sophistication in terms of data manipulation, but I think it's still al long way off being truly semantic from the machine point of view.
So the machine knows Jane is a freid of Mac - does it know what Jane is? A girl? What's a girl - see where this is going?
SEOWorkbench 2 years ago
Nice video, easy and helpful
abolemar 2 years ago
nice, I'm not a programmer, but the subject is really interesting for me
espiriguidiberto 2 years ago
Great tutorial, thanks.
mical33l 2 years ago
Thanx, this is a good tutorial. your style is impressive.
rajshas 2 years ago
Hey This is Great.Thanx a lot.The Way you r presenting is great and easy to grasp.Pls upload more on this area.Any Video About Meta Data Generation?
Any Way Thanx a Lot!
kts1999 3 years ago 2
Great video! The first half is especially nice, bravo!
Please consider using foaf:name instead of <foaf:name> in your N3 example.
Don't you have a feeling that putting semantics into HTML is just as wrong, or maybe even worse than mixing style and structure?
apohoyda 4 years ago 2
Thanks for the foaf:name correction - I'll make sure to make the change it in future videos.
As for the mixing semantics/syntax comment, I think practicality plays a large role in RDFa. There is benefit for keeping the machine readable and the human readable data in sync and in the same document, which is what both Microformats and RDFa does.
msporny 4 years ago
But in your example, the human-readable text "Jane is friends with Mac" is not in sync with the machine-readable data "#jane foaf:knows #mac". You can change the text and the data could stay. Do you call this "in sync"? Aren't you asking people to keep text and data "in sync"? Is there a way to generate the text "Jane is friends with Mac" from the data "#jane foaf:knows #mac". It's not obvious from your example.
apohoyda 4 years ago
I also don't think that having more triples will help computers to "understand" anything. More triples will only allow computers to do simple computations on data, but not more.
This could be used by AI in the future, but by itself triples bring no understanding, IMO.
apohoyda 4 years ago
The point I was trying to make is that we have to lay the foundation first, which is exposing triples on web pages. Once those triples are there, there are many, many types of applications we can build based on that foundation. It is analogous to saying that the more people and pages there are on Wikipedia, the more useful that site becomes...
msporny 4 years ago
I think you can think of this as a way for an AI systems to assemble the data on the internet.
AppA 4 years ago
Yay, brilliant!
dunken69 4 years ago 2
Thanks! Remember to tell your friends about the semantic web! :)
msporny 4 years ago
Very nice :) What's your next topic?
modanbri 4 years ago 2
Thanks Dan :)
The next production is probably going to be an advanced RDFa tutorial. Also wanted to do one on the Microformats approach. We'll be doing several on Bitmunk as well...
If you've got any ideas/suggestions of what you'd like to see next, please let me know!
msporny 4 years ago