The Mouse and the Rectangle

http://dustincurtis.com/the_mouse_and_the_rectangle.html

The article talks about a new addition to the pavolovian response experiment. In Pavlov's experiment, he discovered that dogs could be conditioned to salivate when a bell was rung by ringing a bell every time they were fed. The dogs learned that when the bell was rung that they would be fed, and therefore, they began to associate the bell ringing with food being brought

This article talks about a second experiment performed more recently. A mouse is shown a square and a rectangle. If the mouse goes to the rectangle, he is fed, but when he goes to the square, he is not. The mouse thus learns that if he goes to the rectangle, he will be fed. However, when the square is removed and replaced with an even longer rectangle, the mouse actually does not go to the original rectangle. Rather, it goes to the longer rectangle. This leads to an interesting question. Why would the mouse decide to go after a new alternative rather than the original rectangle that he used to go to?

It was decided that the mouse did not pick based on if the shape was rectangular or not, but rather how rectangular the shape was. The longer rectangle was more rectangular - so the mouse assumed that going there would be more beneficial.

I think the overall pavlovian behavior can be simplified down to an information cascade. Because the mouse is continually fed inputs telling it to go to the more rectangular shape, it knows believes that going there will always provide the greater reward. If the researcher suddenly decides to switch it around so that the mouse is rewarded for going to the less rectangular shape, the information cascade is broken, and the mouse will learn to go to the less rectangular shape.

I think it would also be interesting to see what the mouse would do if it was surrounded by a group of mice who have been taught to go to the circular shape. If the mouse separates from the group and still goes to the longer rectangle, it can be concluded that the mouse has made a distinction between 'less rectangular' and 'more rectangular' and does not act based on the actions of the mice around him. However, if this time the mouse decides to move where ever all the other mice go (so, to the circular shape), it can be concluded that the mouse has been affected by an information cascade.
After seeing how all the other mice go to the circular shape, the mouse could have thought that the others know something that he does not. Therefore, based on the characteristics of information cascades, the mouse gave up his own private information because he believes that doing what others are doing is the more rational thing to do in this case. This would be an example of an information cascade's affect on an individual.

That is interesting. I guess more rectangular means less square. I won't if size (total area) has anything to do with it, as well.