langreiter.com plain, simple: comments-2002-05-02

LANGREITER.COM plain, simple

START

INDEX

Donnerstag, 2. Mai 2002

www.tulip-software.org

2002-05-02-tulip

Erik T. Mueller: "E-commerce and web applications are today largely ignorant of world events. This situation could be remedied if news were converted into a machine-understandable form and made available to these applications. This paper presents an artificial intelligence program called NewsExtract that uses information extraction techniques to transform text news stories into XML NewsForms that represent the key points of 17 types of news events. We discuss the benefits of NewsForms in e-commerce and web applications."

www.panix.com/~erik/pubs/newsund1.htm

2002-05-02-muNews

funzel 8029 days AGO
Nett die ganze REST debatte, obwohl ich REST noch nicht verstanden habe. Wieder jemand, der feststellt dass Dave kein interesse an diskussionen hat sondern nur daran, seine meinung zu verbreiten :-)

David Ness 8029 days AGO
Mueller's paper strikes me as a few good ideas carefully hidden in a pile of real junk. Indeed the examples he uses to introduce the idea of `aware' applications seem so absurd on the surface, to me at least, that I might use them as a parody.

It reminds me of a famous performance by Ed Fredkin (It was a long time ago and it could have been Seymour Papert, but I think it was Ed) on a national, popular TV show back in the 1960s. He was invited to explain `Artificial Intelligence'. Remember, back then computers weren't often discussed by `the public' so this was a real chance to `get the word out'.

Instead, he got going on an example of a device which would run around your head and give you a perfect haircut. Yes, the first haircut would cost thousands of dollars, he admitted, but thereafter they would be quite cheap. In any case the example was so ridiculous that the incredulous host (a) could hardly keep a straight face (this was proving MIT professors were, in fact, just as crazy as the public thought them to be); and (b) never got around to _anything_ else. Gave AI a bad name in the public domain for years, if not decades. And, of course, the idea _was_ pretty ridiculous, in and of itself. In the thirty or forty years since we have made some advances in computers, but still nothing along the line he suggested.

Mueller's paper is squarely in this tradition.

gavin 8029 days AGO
Maybe Mueller could get together with the good people over at Human Markup.org.

HumanML is set forth to be an XML Schema and RDF Schema specification, containing sets of modules which frame and embed contextual human characteristics including physical, cultural, social, kinesic, psychological, and intentional features within conveyed information. Other efforts within the scope of the HumanMarkup TC include messaging, style, alternate schemas, constraint mechanisms, object models, and repository systems, which will address the overall concerns of both representing and amalgamating human information within data.

Perfect!

(I still can't believe HumanML is taken seriously.)

chris 8028 days AGO
I found both your Ed Fredkin recollection and Mueller's proposal very amusing. I guess that Mr Mueller's motivation to explore this field is pretty obvious considering his background as professional pattern matcher. Isn't it primarily a somewhat esoteric extension of the field of technical analysis?

Actually, I'd be very interested in results. What use can computers make of news? Are there already bots that scour the net (stock market commentary etc.) in order to ... at a basic level, make money? I'm sure there are. The question is, do they work (well enough)?

earl 8028 days AGO
the stock market, as always ;)

sometimes it seems to me that the world is just a huge stock market.

chris 8028 days AGO
wie gesagt: sell MSFT at $150.

earl 8028 days AGO
*hrmpf* :)

David Ness 8027 days AGO
I am indebted to Gavin for his reference to HumanML. His remarks caused me to take a look, and I not only agree with his general view, digging around in the site provided some wonderful material for later discussion. Indeed, some of the stuff makes an absolute playground for parody---it seems sufficiently absurd.

As to Chris' question, some work is beginning to circulate. You might want to take a look at my colleague (and frequent collaborator) Steven Kimbrough who not only writes about this problem in general, but has developed some of it sufficiently well to be marketing it.

Please log in (you may want to register first) to post comments!