Before the Fractal Realizer can be defensibly used as a neutral model pattern generator, the synthetic landscapes that it produces must be tested. Although there are scores of landscape pattern metrics and indices, and a number of spatial statistics which could be generated to compare empirical landscapes with synthetic realizations, there is no clear consensus about which metrics should be used, or how closely such spatial indices or pattern statistics must match. Therefore, in addition to comparative pattern analysis tools, we will borrow a higher-order test on which we hope there can be more widespread agreement.
This test was proposed by Alan M. Turing (1912-1954) as a way to determine if a computer posessed intelligence. The test was first described in Turing's 1950 article ``Computing Machinery and Intelligence'' (Mind, Vol. 59, No. 236, pp. 433-460).
In the Turing Test, an interrogator is connected to a human and a machine via a terminal, and therefore can't see his/her counterparts. The interrogator's task is to determine which of the two candidates is the machine and which is human exclusively on the basis of their responses to questions. If the interrogator is unable to correctly identify the human subject within a certain time (Turing proposed five minutes, but the exact amount of time is generally considered irrelevant), the machine is intelligent.
The human eye and brain are excellent at pattern integration and recognition. If human experts cannot distinguish an actual landscape from a synthetic landscape created by the Realizer, then the Realizer may have passed the most stringent test of all.
You are invited to participate in the Turing Test of the Realizer. You will be presented with a series of 20 selections of paired maps, and asked to distinguish the map which is real from the synthetic realization.
One member of each pair of maps is real and has been clipped from an actual map. The other is synthetic, and has been generated using the Fractal Realizer. Please select the landscape map that you think is real.
You are to make your selections on the basis of the patterns alone; therefore, no information about the empirical maps will be provided before the test about land cover types, scale, etc. The map colors are earth tones chosen to avoid the suggestion of particular actual land uses or cover types.
The synthetic landscape realization from each pair is generated by the Fractal Realizer ``on the fly'' while you take the Test; therefore, no two people will experience exactly the same Test.
After the Test, you will be told how many times you correctly identified the real map. Turing Test results from nearly 40 participants will be compiled and presented (anonymously) as a data table in a scientific publication about the Fractal Realizer.
The Turing Test takes about eight minutes, assuming that you spend an average of 20 seconds to make your selection from each pair of maps.
Take the Turing Test of the Fractal Realizer
For additional information contact: