Now, as regards e.g.

“But what if it is not even necessary to build the first generation of semantic tools? What if instead of trying to teach computers natural language, we hard-wired into computers the concepts of everyday things like books, music, movies, restaurants, stocks and even people. Would that help us be more productive and find things faster?”

The answer is -perhaps- that if you actually did hard-wire into computers concepts of all these “everyday things, music, etc”, what you would effectively end up doing is to create a valid Ontology-of-everything, akin to the famous (now open-source) CYC ontology, with an incredible number of semantic and logical definitions already built-into it, yet (according to many people) STILL not enough.

The idea is not wrong, nor unfeasible; on the contrary, it’s a logical step to take. Moreover, you really don’t have to use such a huge-size Semantic structure, for a good top-down project like those you propose; all you need to do is identify the problem-specific ontologies and semantic information that is relevant to the problem, as well as write special tools, i.e. natural-language understanding tools with a limited scope and limited parsing ability, to process all those dumb web-pages at the bottom-end with the wisdom of your top-end application(s)…

I.e. what you say is not very different from what certain projects based on NLP are already doing (e.g. POWERSET).

I collaborate with a company who need A.S.A.P. a semantic application that works with existing webpages, without RDF or other bottom-up semantic stuff, so I understand what you are saying; this is what we already do (given the lack of something else - more powerful).

However, after more than two decades in NLP and Prolog I am one of these A.I.-sympathisers who feel that the best solution -ultimately- will be the use of inferences and NLP at all levels, both top-down and bottom-up. It will take time, but we can already achieve a lot with what we have; it’s already quite powerful…

Every economic revolution has its ups and downs, but more opportunities are now available for young up-starts to get a jump on the industry of their choice than at any other time in history. Decentralized businesses are booming; no longer are the thinkers and dreamers with the best ideas stifled, they are free to develop (and profit) to the absolute extent of their abilities.

Granted, manipulating web 2.0 and successfully utilizing recent advancements in social networking, all while dealing with the prominence of trans-national corporations requires new skills, but as humans, we display a remarkable tendency to adapt to whatever challenges lie before us — and I expect a continuing trend of success.

Generally speaking, this is the great step that could completely change our world as far as we know it now, like when computers were created and developed or just like Google, and it is all this tiny but huge things that have been changing our way of living and the way we understand life.
This is an open question, and I want that it would be the spirit of this simple question! You are all invited to build the meta internet! Then you could start thinking a lot about this issue!

——————————————————————————-

From:
Francisco A Ceron G /Seeking for a new Job! Top 2,6K+ 8,6M+

You say: “it would be interesting to have some insight and analysis of what the boundary actually is”, and I ask you, Where do yo start the insignt and the analysis of what the boundary actually is? Where is the rigth starting point?.

Best Regards, Francisco

On 1/27/08 5:13 AM, Oren Yosifon wrote:
——————–
Hi,
I think you are taking some of the metaphors used to describe the vision of the Semantic-Web and actually attribute them to Semantic-Web-Technologies. By no means these technologies should be considered as a replacement for human thinking.

If you looked deeper into some of the specifications, you could have seen that the documentation quite agrees with you: it is taken as a known-fact that higher levels of formalisms and rich-semantics that go beyond Description-Logic may be computationally unfeasable. This is why, for example, there is a difference between OWL-Lite, OWL-DL and OWL-FULL, where the Lite and DL Versions are guaranteed to have reasoners operate over them in finite time.

Ontologies, RDF, OWL, Triple-Stores, Inference-Engines and the likes, are used in production environments to accomplish tasks such as information-integration , text-annotation and information-retrieval .

So, I think instead of proving the already-known (Semantic-Web is limited) it would be interesting to have some insight and analysis of what the boundary actually is.

Best Regards,
OrenYosifon

- Steven Burda
http://www.linkedin.com/in/burda
burda.mba[at]gmail.com

Let’s suppose that people actually manage to laborously complete the whole processing of tagging and attributing through XML and actually give meaning to data. So according to the bottom-up Semantic Web approach, the Common Brain will give us answers and we’ll take it, right?

But 3 problems I see in this:
first is that the beauty of tagging in web 2.0 is all about how it’s individual and NOT conglomerated into a greater whole; people decide whether they wanna file the news about Lauren Hill shaving her head under hair fashion or under Jewish artists or whatever, much like how you’ve filed this article under many tags. So what’s the fun if someone else does it for you, so authoritatively?
second is, what;s the point if we do all the manual work and leave the processing to machines? What’s the difference between web 2.0 and web 3.0? that there’s application to all the manual labor?
the third is, on extension of the forementioned two and my initial concern - where does our own skills of info processing and research go? in other words, and already mentioned before, how do we know what’s good info and bad? Do the machines determine it for us?

Well clearly this bottom to top approach doesn’t have any AI built into it save for the processing part, so my REAL concern is that info processing and its categorization (and even more, management) will be in hands of the initial coders, so to say. And since this coding is not as public and open-to-all-free like web 2.0, I fear the coming of lot of unequal approach (and rendering) of info and data, in this “egalitarian” age of pursuing all and equal access.

Some may argue that top-down approach technology (i.e. Google’s and FOAF data(http://noisebetweenstations.com/personal/weblogs/?p=2147) and Linguistic Agents (www.linguisticagents.com) may be more authoritive since there’s not much room for human tagging and editing, but my personal opinion is that these top down approach at least separates itself from the human approach - it only assists the search, not give “the answer” to the search.

Thirdly User before provider. Process before data.

The last (third is primarily) dictated by a marketing approach any way. As to the second all worthwhile services would address the composition perhaps in ad-hoc fashion.

Question is How much of developer time is saved? How much more efficent is the new proposed mechanism? Any bench marks available?