Defined as a Computational Knowledge Engine. It does an awful lot of number-crunching but looks more as a giant closed database than a distributed Web of data, or even a ‘Semantic web’.

Interesting the reaction of competitor Doug Lenat, who although is claiming that Wolfram is not AI (therefore CYC’s got nothing to fear) imho is realizing he didn’t get it right when he set out trying to capture ‘common sense’. After all, all that we find in Wolfram|Alpha is symbolic reasoning. Is that so far from the way Cyc works? This might be a nice departure point for an interesting discussion…

Lenat’s blog post contains some interesting comments on the things that Wolfram|Alpha can’t do (yet):

When it returns information, how much does it actually “understand” of what it’s displaying to you?  There are two sorts of queries not (yet) handled: those where the data falls outside the mosaic I sketched above — such as:  When is the first day of Summer in Sydney this year?  Do Muslims believe that Mohammed was divine?  Who did Hezbollah take prisoner on April 18, 1987? Which animals have fingers? — and those where the query requires logically reasoning out a way to combine (logically or arithmetically combine) two or more pieces of information which the system can individually fetch for you.  One example of this is: “How old was Obama when Mitterrand was elected president of France?”  It can tell you demographic information about Obama, if you ask, and it can tell you information about Mitterrand (including his ruleStartDate), but doesn’t make or execute the plan to calculate a person’s age on a certain date given his birth date, which is what is being asked for in this query.  If it knows that exactly 17 people were killed in a certain attack, and if it also knows that 17 American soldiers were killed in that attack, it doesn’t return that attack if you ask for ones in which there were no civilian casualties, or only American casualties.  It doesn’t perform that sort of deduction.  If you ask “How fast does hair grow?”, it can’t parse or answer that query.  But if you type in a speed, say “10cm/year”, it gives you a long and quite interesting list of things that happen at about that speed, involving glaciers melting, tectonic shift, and… hair growing.

Some nice coverage of it also on ReadWriteWeb and UMBC.

…

The seminar’s title is “reasoning about knowledge‘. Fagin and others, lead authors in the area, define its scope as follows (get the PDF here):

As its title suggests, this book investigates reasoning about knowledge, in particular, reasoning about the knowledge of agents who reason about the world and each otherâ€s knowledge. This is the type of reasoning one often sees in puzzles or Sherlock Holmes mysteries, where we might have reasoning such as this:
If Alice knew that Bob knew that Charlie was wearing a red shirt, then Alice would have known that Bob would have known that Charlie couldnâ€t have been in the pantry at midnight. But Alice didnâ€t know this . . .
As we shall see, this type of reasoning is also important in a surprising number of other contexts. Researchers in a wide variety of disciplines, from philosophy to economics to cryptography, have all found that issues involving agents reasoning about other agents†knowledge are of great relevance to them. We attempt to provide here a framework for understanding and analyzing reasoning about knowledge that is intuitive, mathematically well founded, useful in practice, and widely applicable.

For the moment we’ve just been clarifying our language and the conceptual tools we need to move on to the core issues. But the discussion’s been really lively, so I guess I’ll keep posting about this. A simple map of the recent meeting is online for public consumption :-)

I guess that many of the people originally supporting the SW vision didn’t really have the time to worry about this sort of questions, as they had different background, or maybe were just so excited about the grandiose idea an intelligent world wide web interconnected at the data level. Quite understandable, but as the idea is now reaching out to the larger public and maybe connecting to the more bottom-up Web2.0 movement, I think that it’d be great to re-think the foundations of the initial vision. Also with some rigorous clarification about the terms we use. The article of Chiara Carlino reaches an interesting conclusion:

So-called semantic web technologies provide the machine with data, like chinese symbols, and with a detailed set of instructions for handling them, in the form of ontologies. The computer simply follows the instructions, as the person in the chinese room does, and returns us useful informations, avoiding us the task of processing a big set of data on our own. These technologies have in fact nothing to do with semantics, because they never refer to anything in the real world: they never have any meaning, except in the mind of those expressing their knowledge in a machine-readable language, the mind of those preparing chinese symbols for the person in the chinese room. The person in the room – the machine – never ever gets this meaning. Such technologies, eventually, deal not much with semantics, but with knowledge, and its automatic processing through informatics. It seems therefore misleading and unfitting to keep on pointing with the word semantic a not semantic at all technology. It looks quite necessary to find out a new term, capable of hitting the core of this technology without giving rise to misunderstandings.

The article was also posted on the w3c SW mailing list some time ago, and generated an interesting discussion. But then, if we have to throw away the overrated ‘semantic web’ term, how should we call it instead? Without any doubt, this research strand has generated lots of interesting results, both theoretical and practical. Mmm maybe, mainly practical – see the many prototypes, ontologies and standards for manipulating ‘knowledge’. So, continues the author, what people are doing is not really dealing with ‘semantics’, but building very complex systems and infrastructures for dealing with ‘knowledge structures’:

There is a word who seems to serve this purpose, and that is epistematics. Its root – epistéme – points out its strict connection with knowledge; nonetheless, it is not a theoric study, not an epistemology: it is rather an automatic processing of knowledge. The term informatic has been created to point out the automatic processing of informations: similarly the term epistematic is pretty fitting in pointing out the automatic processing of knowledge that the technologies we are speaking about make possible. The terms also reminds informatics, and this is pretty fitting as well, as this processing happens thanks to informatics. Eventually, the current – though not much used – meaning of epistematic is perfectly coherent with the technologies weâ€d like to point out with it: epistematic, in fact, means deductive, and one of the most advanced features of these technologies is exactly the chance to process knowledge deductively, using automatic reasoners who build into software the deductive rules of formal logics. The formerly so-called semantic web looks now like a new science, not bounded (and narrowed) anymore to the world of web, as the semantic web term suggested: epistematics is a real evolution of informatics, evolving from raw informations processing to structured knowledge processing. Epistematic technologies are those technologies allowing the automatic processing, performed through informatic instruments, of knowledge, expressed in a machine- accessible language, so that the machine can process it, according to a subset of first order logic rules, and thus extract new knowledge.

I like the term epistematics – and even more I like the fact that the ‘web’ is just a possible extension to it, not a core part of its meaning. Semantic technologies, based on various groundbreaking works the AI pioneers did some twenty or thirty years ago (mainly, in knowledge representation), have been used much before the web. Now, is the advent of the web making such a big difference to them? They used to write knowledge-based systems in KIF – now they do them in OWL – we change the language but aren’t the functionalities we are looking for the same? They used to harvest big companies’ databases and intranets to build up a knowledge base – now we also harvest the web – is that enough to claim the emergence of a new science, with new problems and methods? Or is it maybe just a different application of a well-known technology?

I must confess, the more I think about such issues, the more I feel they’re difficult and intricate. For sure the web is evolving fast – and the amount of available structured information is evolving fast too. Making sense of all this requires a huge amount of clarity of thought. And presumably, this clarity of thought will eventually drive to some clarity of expression. Wittgenstein wasn’t the first one claiming it, but for sure he did it well: language plays tricks on us. Better, with his words:

Philosophy is a battle against the bewitchment of our intelligence by means of language.