Primal is a semantic engine

Blog

6 Responses to “Thought Networks Don't Need to Socialize”

Thanks to DEMO and Nova Spivack who assembled and moderated the panel. The panellists, Jon and Howard, Peter Norvig from Google and Prabhakar Raghavan from Yahoo, provided an engaging discussion. I highly recommend investing some time with it. The portion I’ve described begins around the 22-minute mark. The link to the video is in the opening paragraph of this post.

And for more discussion between Jon and Howard on “The Global Brain”, check out this edition of Interviews with Innovators.

In a nutshell, I would simply extend Jon’s comments by saying: we derive even more knowledge from the combination of links between our documents, and the links between entities within our documents, across our private and public data spaces (e.g. weblogs, wikis, shared bookmarks etc.).

Basically, we are making “What” connections across a Linked Data mesh (structured data network) that starts from the “Who” node in a semi-structured social-network.

The evolution of networking via the different layers of abstraction from router focus, to computer name focus (internet), to current document url focus (document web), and eventual entity URI focus (linked data web) is something I cover in a remix of my Linked Data Planet confderence keynote. The remix simply pulls in stuff from presentations TimBL has given about the same mater in the past re. the Semantic Web etc..

Links:

1. http://tinyurl.com/4pemgk – follow slides 15 – 22
2 http://ode.openlinksw.com – if you want to see this in practice on the Web today (start with the presentation since it is RDFa based)

Kingsley

The title “Thought Networks Don’t Need to Socialize” brings up a very interesting message to the person who views the described panel from the science history PoV; like what does this claim “Thought Networks Don’t Need to Socialize” mean in a bigger picture?
It means that the science (OK, the discipline) that deals with the problem has entered the stage of narrowing the research area, the stage of further and deeper defining its scientific object/subject.
If we have a look at how humanity developed its knowledge in various fields, we might notice that the object in focus was blurry at first and gradually became better defined and described.
Hence my suggestion: let us take advantage of the moment and clear out the picture; to do that we can ask the researchers to write the lists of:
1) what the thought network is NOT;
2) which features we notice (like, prefer, appreciate) the most;
3) what does it look like?
4) what are its manifestations?
5) which features are ‘bad’ (not desirable)?
6) which desirable features seem to be missing?
And then call for a brainstorming round table to share the ideas inspired by the lists.
That discussion might end up in developing a new research programme based on a more clear idea of where we are going.

Peter, thanks for this post: it makes it quite clear that nobody really wants to dwell into the peculiarities of something called “ontology”, which is Ok with me; the need to go there during a decision making process arises rarely, only when everything else stalls. At this moment your account of the current situation tells me that much:
– the actors in the field are of the same ilk and circumstances and they understand each other pretty well (one more reason not to bother with ontologies);
– we are talking of a new generation for an internet search engine;
– probably that will demand a new websystem, some more additional software (that will help to convert data from a format into any other);
– the most concern is to make it “smart”, similar to what Orson Scott Card described in his award-winning “Speaker for the Dead” as Jane. This Person-Device constantly sifts through all the possible data in the universe and gives Andrew Wiggin all the information he needs for every specific question that he poses in the format that suits most for that particular case. It is like a Jinn who is always ready to help you under any circumstances, and never acts on its own will (John Brisbin has recently voiced a concern: should AI have a will of its own. See his comment earlier in this blog.

However magic this sounds, it gives us an idea, a concept of what AI might look like.
Does it make any sense?

You want to know that it is?

Luckily, this is already done and shipped to millions of users, we don’t need to wait for Primal Fusion to tell us the obvious.

check out nepomuk.kde.org

The Semantic Desktop initiative followed this goal for many years now.

I think all would agree there’s a gap between our technical capabilities and the popular understanding of those capabilities. To argue what’s possible is not to argue what’s perceived as possible. And at the moment, it’s difficult to imagine these semantic networks existing independently from the document and social networks that are so dominant today. This is a constraint of our collective imagination, not our technology. The products we’re inventing use “semantic” as the qualifier for these primary networks/media: semantic PIMs, semantic document managers, semantic desktops, semantic social networks, semantic search, etc. The different types of networks obviously benefit from integration and cross-pollination, but semantic networks deserve a life of their own.

Leave a Reply