Stochastic General Recognition Theory

随机一般认知理论是知觉和认知心理学中流行多年的几个概念的精确数学形式化。其中之一是，反复接触任何复杂的刺激并不总是产生相同的感知状态；第二，分析这种刺激的感知系统由几个相互作用的通道组成，这些通道并行运行。该理论在广泛的认知和知觉实验中对反应概率和反应时间做出了具体的数字预测，例如识别和分类。此外，它还有可能解释不同刺激暴露时间的影响，一个刺激与另一个刺激的掩蔽，以及短刺激间隔的影响。这项研究将包括一些实验，这些实验的结果是由理论预测的；因此，这些实验将为理论提供严格的测试。一组实验将考察一个类别的结构如何影响分类时间；另一组将考察不同刺激成分的知觉加工过程中的不同类型的交互作用如何影响分类和识别性能。这些实验将增加我们对人类如何识别和分类复杂模式的理解。反过来，这可能会提高我们对一些阅读和学习障碍的理解，还可能导致机器更准确的模式识别。

Stochastic general recognition theory is a precise mathematical formalization of several ideas that have been popular in perceptual and cognitive psychology for many years. One of these is that repeated exposure to any complex stimulus does not always elicit the same perceptual state; a second is that the perceptual system that analyzes such a stimulus is composed of several mutually interacting channels that operate in parallel. The theory makes specific numerical predictions about response probabilities and response times in a wide variety of cognitive and perceptual experiments, such as identification and categorization. In addition, it has the potential to account for the effects of varying stimulus exposure duration, the masking of one stimulus with another, and the effects of short interstimulus intervals. This research will include a number of experiments whose results are predicted by the theory; thus, these experiments will provide a rigorous test of the theory. One set of experiments will investigate how the structure of a category affects categorization time; another set will investigate how different types of interactions during the perceptual processing of separate stimulus components affect categorization and identification performance. These experiments will increase our understanding of how humans identify and categorize complex patterns. This, in turn, may improve our understanding of some reading and learning disabilities and may also lead to more accurate pattern recognition by machine.