Freebase: the Web 3.0 machine
March 09, 2007
Artificial intelligence guru Danny Hillis has launched an early version of the first major Web 3.0 application. It's called Freebase, and its grandiose epistemological mission is right up there with those of Google and Wikipedia."We’re trying," Hillis tells John Markoff of the New York Times, "to create the world’s database, with all of the world’s information.” Alpha user Tim O'Reilly says that Freebase "appears to be a bastard child of wikipedia and the Open Directory Project" but that it's really "like a system for building the synapses for the global brain.”
The product of Hillis's latest company, Metaweb Technologies, Freebase is a user-generated brain. Like Wikipedia, it allows people to freely add information to it, in the form of text or images or, one assumes, anything else that can be rendered digitally. But it also allows users to add "metadata" about the information - tags that describe what a word or picture is and how it relates to other information. Freebase, says O'Reilly, "turns its users loose on not just adding more data items but making connections between them by filling out meta tags that categorize or otherwise connect the data items, using a typology that can be extended by users, wiki-style."
The addition of rich meta tags in a standardized form is what makes Freebase a next-generation Web application - a manifestation of what Tim Berners-Lee long ago dubbed the Semantic Web and what has recently been rebranded Web 3.0 for popular consumption.
Although the wikipediaesque user-generated quality of Freebase will get much attention, Freebase is really more about the creation of a community of machines than a community of people. The essence of the Semantic Web is the development of a language through which computers can share meaning and hence operate at a higher, more human level of intelligence. The meta tags are crucial to that machine language. Freebase hopes to harness the (free) labor of a big pool of vounteers to add those tags, which is a labor-intensive chore (and a big hurdle on the path to Web 3.0).
Should Freebase pan out - and right now it's largely a theoretical construct - it would have many practical (and money-making) applications. It would provide the basis for a more natural form of searching, allowing programmers, as Markoff says, "to write programs allowing Internet users to pose queries that might produce a simple, useful answer rather than a long list of documents." It would also enable various information-processing devices that used to have to be configured manually (by people) to be able to program themselves automatically. A rudimentary example is "the video recorder of the future," which "might stop blinking and program itself without confounding its owner."
But Hillis has bigger fish to fry than self-programming gadgets. In the past, he's expressed a desire to create machines that transcend what he sees as the limitations of human beings. "I guess I'm not overly perturbed by the prospect that there might be something better than us that might replace us," he once said. "We've got a lot of bugs, sorts of bugs left over history back from when we were animals." Freebase is an attempt at creating an artificial intelligence that can be bootstrapped by the contributions of humans. On one level, it works for us. On a deeper level, we work for it. As Hillis has also said, Web 3.0 is a "spooky thing."
Of course, relying on a rag-tag band of volunteers, all afflicted with those nasty evolutionary bugs, brings its own problems, particularly in an effort that, unlike Wikipedia, requires a great deal of consistency and precision in terminology. Freebase's ability to attract and manage a human horde will be critical to its success. Will we be up for the job?
What if this didn't turn out to be a precise and complete body of work? It will still serve a useful (albeit a limited one) purpose for its creators & users. Along with it there would be other ontologies to complement the limitations of this one, effectively creating a semantic babel, at surface chaotic, yet enabling machines/humans to exchange knowledge. Why? because humans have figured out how to navigate through language, cultural, and other barriers to exchange information/knowledge. This is a relatively easy one to surmount.
Please note that DBpedia (http://dbpedia.org) has created a bona fide Web 3.0 Database from Wikipedia. Of course, there are many of these to come as part of the regalvanized "Open Data" movement (see: http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData ).
BTW - why can't I use my OpenID for authenticating my posts? Typkey is OpenID based so any OpenID account will/should work here :-)
Posted by: Kingsley Idehen at March 9, 2007 04:26 PM
¿HAL 9000 en camino?
El sueño de Danny Hillis hecho realidad... Bueno, digamos que una invitación a todos los mortales para que contribuyamos a él (al sueño)... pues seguiremos siendo mortales ¡pero Freebase no!
Este amigo Danny Hillis se escribió por allá a finales del siglo pasado en pleno apogeo de la primera ola de la Web un artículo en Wired que por lo menos al suscrito causó tremenda impresión (verlo aquí: http://www.wired.com/wired/archive/6.01/hillis_pr.html)
Ahora creo que no estaba escribiendo en broma y que además decidió pasar de la reflexión a la acción ¡GUAU! que en verdad lo que podría significar la creación de esta "máquina" de los significados no es poca cosa... Habrá que ver qué nace de todo esto: si lo que dijo Kant "del árbol torcido de la humanidad no puede resultar nada a derechas" (un HAL 9000), o, si somos capaces, de alguna forma, de reinventarnos, como por ejemplo lo (d)escribió Michael H. (ver aquí: http://rpizarroeu.blogspot.com/2007/02/la-s-era-s-del-hielo-suramrica-y-el.html)
Los avances de la tecnología en toda época nos hacen exclamar ¡Que décadas tan interesantes estas que nos están tocando vivir!
NOTA 1: Hal 9000 es un "PERSONAJE" de novela, ¿Recuerdan 2001 Odisea del Espacio?, ¡Y qué personaje (para ser una máquina)!, Arthur C. Clarke, escritor, Stanley Kubrick, director de la película
NOTA 2: HAL 9000 viene del acrónimo inglés Heuristically programmed ALgorithmic computer (Computador algorítmico heurísticamente programado)
It sounds very similar to nndb
Posted by: Rose Water at March 12, 2007 06:03 PM
I agree with your analysis. In the end, the metadata tags themselves will have to be semantically consistent. To my knowledge, that is a challenge.
I did signup to test drive the system. The usability and appeal of end-user tools to my mind will make or break the effort.
Posted by: Arthgallo Wachs at March 14, 2007 11:42 PM
I have developed my own definition of Web 3.0, and differ on the viewpoint that Semantic Web would be the essence for the next generation of the Internet.
And for a series of examples, you can see my analysis of the Personal Finance category from a Web 3.0 perspective.
Posted by: Sramana Mitra at March 27, 2007 08:03 PM
Post a comment
Thanks for signing in, . Now you can comment. (sign out)(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)
"Riveting" -San Francisco Chronicle
"Rewarding" -Financial Times
"Ominously prescient" -Kirkus Reviews
"Riveting stuff" -New York Post