Daily crono / The creativity article

Daily crono

No problems of free time today, either. Waking up at 7, quick breakfast, walking to MIT with Mr Kumar, 8-9.30 RFID class ( a person from ThingMagic explaining the physics of RFID and giving some demonstrations, pretty interesting), 9.30-10 Master Thesis seminar (on simulations and systems engineering, quite OK, but I joined in the middle, but made a comment still, or two), 10-12 writing my thesis proposal, 12-13 lunch in the local Bengali restaurant (good food, and plenty of salad), 13-14.30 reading Utterback's books for the disruptive technology class, 14.30-16 disruptive technology class (interesting presentation about Lotka-Volterra models by Moise Solomon, MITRE. The model presents quite accurately how different technologies develop their market-share, more about this later, perhaps), 16.30-18 an interesting lecture about product architecture by professor Whitney, 18-19 talking within the group about product development assignment, 19-20 writing interview plans in a library and reading an interesting article about generating ideas using computers (someone has implemented a program, which produces better advertisements that the ADs do, but IMHO it is just showing that computers are good in mechnanical task, i.e. producing lots of combinations from existing lists of ideas, nothing deep in this, but interesting still), 20-21 participating in Chinese new year celebrations (free food and beer, and typical Chinese program), walking back home, finalizing the book report, blogging, and wrapping up the day by reading some articles and books, getting to bed at 23.30 or so. Quite much in one day.

Oh, I even spent 10 minutes reading New York Times and got to know that e.g. Lance will race in Tour De France, and that mr Greenspan supports mr Bush' retirement system privatization even though it will result in borrowing trillions of dollars. And now I will stop blogging for today.

The creativity article

The main claim in the article is that the so called random theories of creativity are wrong. Those theories (like brainstorming, random input etc) claim, that best ideas are found by creative sparks when the normal rules and boundaries of thinking are cast aside. According to the article, there have been and will be such sparks, but they are rather rare. For most of the time, we would be better by just following some field specific templates. The article bases this claim on extensive literature survey, and on the experiment of using computers to create new advertisements. The methos was as follows: take a product and some of its traits (like softness, safety etc). Find another product/symbol, which also has the same trait. Replace the second product with the first. E.g. like bullet shaped car to denote a fast car etc. The authors collected a few hundred such relationships from magazines, and then used computer the combine them in novel ways (more or less enumerating all possibilities). Then they asked laymen to create advertisements by using similar template technique and without using such a technique. Then some judges rated the results. In all cases, the computer generated advertisements were considered to be more creative by the judges. So, does this prove anything? I am not sure. I tend to agree (based on my experience) that random shooting in creating ideas is not the most efficient way to be creative: one usually does better by following some patterns (adding some formalism to brute force). One certainly cannot create the most creative scientific theories by computer, not can write novels automatically, I think. But using computers might help in creating e.g. system architectures, and other somehow limited artifact, at least in a given, well known application domain. Well, this is basically what Mr Koo did in his thesis. It is also what someone did using genetic algorithms: he created some at one time novel electrical circuits automatically using genetic algorithms and some well know goodness metrics. If this is not enough, read the article "Templates in Creative Sparks".

Alun perin julkaistu kuukausikoosteessa "Blogs February 2005".