Letter perception emerges from unsupervised deep learning and recycling of natural image features

[ ]

There is a recent paper on Nature on unsupervised deep learning of recognizing letter images.


The use of written symbols is a major achievement of human cultural evolution. However, how abstract letter representations might be learned from vision is still an unsolved problem1,2. Here, we present a large-scale computational model of letter recognition based on deep neural networks3,4, which develops a hierarchy of increasingly more complex internal representations in a completely unsupervised way by fitting a probabilistic, generative model to the visual input5,6. In line with the hypothesis that learning written symbols partially recycles pre-existing neuronal circuits for object recognition7, earlier processing levels in the model exploit domain-general visual features learned from natural images, while domain-specific features emerge in upstream neurons following exposure to printed letters. We show that these high-level representations can be easily mapped to letter identities even for noise-degraded images, producing accurate simulations of a broad range of empirical findings on letter perception in human observers. Our model shows that by reusing natural visual primitives, learning written symbols only requires limited, domain-specific tuning, supporting the hypothesis that their shape has been culturally selected to match the statistical structure of natural environments8.


Deep learning architecture and examples of natural image and printed letter data.


Simulations of human psychophysical studies.

Spatial-frequency analysis of perceptual channel mediating letter identification.

Written on August 22, 2017