Patented computer software may replace human authors
 

Imagine all you had to do is control a computer program, and then the computer does all the title content creation for you. Sounds like the fictional example of an artificial intelligence? Well, Philip M. Parker, Professor at INSEAD University, has created this special-purpose software assistant which automatically produces a set of similar books from a template that is filled with data from database and internet searches. And guess what -- at Amazon, Mr. Parker is currently listed as the author of over 85,000 books!

Check out the video below for a demonstration.



[via Gizmodo, thanks to Madam Broshkina for the tip!]

This is really crazy, but I don't think that the papers the software creates are worth reading. In the end it's a computer and not a human who writes the content...

But anyway, can it be ported to the iLiad? :2thumbsup

You really need to read the NY Times article on this. This guy is just cranking out crap, hoping to make a fast buck. This was also discussed over at TeleRead two days ago.

A link to the original NYT article:

http://www.nytimes.com/2008/04/14/bu...in&oref=slogin

Well, I am not into these type of "books" and I use the term books loosely in this context. Give me an author who really loves to write...

This guy's a piker, at least in pure volume terms. Go to Amazon and search for "Jassen Bowman" and get a look at the results: 143,056+ listings, for audiobooks (on CD) with titles like, Learn the Hair Lotions Online Business Networking Secrets. I'm guessing that all the CDs are either identical or lightly templated, assuming that they exist at all. (Every one I've seen is marked "currently unavailable." The trick here is probably generating all the listings via some kind of bot.

A nice short discussion of this phenomenon is available here.

Parker is mentioned in that same blog entry. To make a living it only takes a couple of "what the hell; I'll try it" sales per title when you have that many posted.

Complaints have been lodged with Amazon since early 2007; one would think that an online retailer of that size and sophistication would be on to them by now.

If Amazon can make a buck off it, what's their motivation to get rid of it? I just purchased something at Amazon that was crap and tried to write a review to warn others, come to find out they block negative reviews (so beware of products with no reviews). They returned my money. They know the item doesn't work. It's still being sold on their site.

If that guy was just out for the quick buck, I wonder how he got to be a teaching professor at a renowned university such as INSEAD.

That must be the software that creates all the Hollywood blockbuster screenplays.

:rofl:

Okay, first I was impressed.
But all of those titles sound like crap, and some of the prices are astronomical!!
795 bucks for "The 2007-2012 World Outlook for Rotary Pumps with Designed Pressure of 101 to 249 P.s.i. and Designed Capacity of 11 to 99 G.p.m."???

What does one have to do with the other? He may be a very good professor. That doesn't stop him from also having a sideline as a shyster. One is not exclusive of the other. Actually, I imagine that he is using the fact that he is a professor to help give respectability to his robo-book venture.

Lets just say I'll feel ripped off if I ever purchased any of his books.

hey wait a sec...

OH MY GOSH!!!!!!!

I have to say that I find the academically aspect of what he did very fascinating. Content creation based on simple facts and resulting in human readable (enjoyable?) sentences is a very difficult topic. Having worked in the artificial department at my University for some time, I'd be very interested in how exactly his logic based synthesizer works.


That being said, I think that the application of this technique he presents in this video is... disturbing.

Automatically creating statistical reports is one thing - I assume that nobody will ever want to read more than one of those in her life. This stuff is boring enough when just written by humans. Having it template-based done by algorithms denies the result everything but the bare facts: not a single bit of experience (which would be applied by the human author), no conclusions, and no unique distinction. The example, given in the NYT article, proves the naivety of the approach: extrapolating data from a related study and filling in the template in question might be much faster for a computer. Though I have reason to believe that an adequate presentation compiled and written by a human experienced in this type of work might not take much more than one day work resulting with information in a more compact and thus consumable form.
Probably would have taken the human even less than thirty seconds to conclude that using antipsychotic drug statistics of the USA are in no way applicable to worldwide use (not even when reducing the area to industry nations only) and therefore that the whole task is moot.

But publishing language dictionaries, thesauri, and medical dictionaries which are (according on user comments) based on free accessible web information is more than a bit dubious (not to say ethically controversial). Sure it wouldn't sell as much if he just declares that he has a decent web crawler, data mining, and linguistic statistics engine and sold the result without the (questionable task of) sentence synthesis.

However creating language education programs or even game show scripts with this technique is just.... dumb. No wonder fewer and fewer people are able to talk in a competent and understandable way nowadays if their learning process is based on statistics where words are taught with no emotional infliction or any other contextual value (and I don't even want to think about the stupidity of teaching language in a shot-em-up game).

But my highlight where the words about his future work. Not that I'm much into romance novels, though I think that there is more to it than "only so many body parts". And the (although ironically meant) comparison of possible advancements in poem creation to Shakespeare cannot be covered with hubris anymore.

In my point of view Mr. Parker lacks the understanding of the distinction between eloquent and good writing. He's seems more the quantity over quality kind of person.

Well honestly I dream myself of an *useful* "paper generator". Like one where you enter as keywords for example "bugs" and "vulcan", and will write an essay from an information base with all it knows about what bugs have to do with vulcans...

Otherwise a paper generator is nothing new:

http://pdos.csail.mit.edu/scigen/

This outputs already nice academic papers, that without some reasonable will to actually read it you hardly can tell it to be complete random bullshit.

This whole story remembers me somehow about *Racter* and "the policemans beard is half-constructed", that was an enourmous hype and made his creator famous, when later the software was analysed, it turned out to be quite a fraud. The software didn't make any texts, it only was available to mangle existing texts, until they appear to be something original... and a decade later, all the templates analysed Racter could never have created a sentance like "the policemans beard is half-constructed"...