Probabilistic principles in unsupervised learning of visual structure: human data and a model

To find out how the representations of structured visual objects depend on the co-occurrence statistics of their constituents, we exposed subjects to a set of composite images with tight control exerted over (1) the conditional probabilities of the constituent fragments, and (2) the value of Barlow's criterion of "suspicious coincidence" (the ratio of joint probability to the product of marginals). We then compared the part verification response times for various probe/target combinations before and after the exposure. For composite probes, the speedup was much larger for targets that contained pairs of fragments perfectly predictive of each other, compared to those that did not. This effect was modulated by the significance of their co-occurrence as estimated by Barlow's criterion. For lone-fragment probes, the speedup in all conditions was generally lower than for composites. These results shed light on the brain's strategies for unsupervised acquisition of structural information in vision.

[1]  R. Aslin,et al.  PSYCHOLOGICAL SCIENCE Research Article UNSUPERVISED STATISTICAL LEARNING OF HIGHER-ORDER SPATIAL STRUCTURES FROM VISUAL SCENES , 2022 .

[2]  M. Goldsmith,et al.  Statistical Learning by 8-Month-Old Infants , 1996 .

[3]  Linda B. Smith,et al.  The importance of shape in early lexical learning , 1988 .

[4]  Geoffrey E. Hinton,et al.  Learning Population Codes by Minimizing Description Length , 1993, Neural Computation.

[5]  S. Palmer.  Hierarchical structure in perceptual representation , 1977, Cognitive Psychology.

[6]  Elie Bienenstock,et al.  Compositionality, MDL Priors, and Object Recognition , 1996, NIPS.

[7]  Nathan Intrator,et al.  Blurred face recognition via a hybrid network architecture , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.

[8]  Scott P. Johnson,et al.  Visual statistical learning in infancy: evidence for a domain general learning mechanism , 2002, Cognition.

[9]  S. Edelman.  Constraints on the nature of the neural representation of the visual world , 2001 .

[10]  Nathan Intrator,et al.  (coarse Coding of Shape Fragments) (retinotopy) Representation of Structure , 2000 .

[11]  Alex Martin,et al.  Properties and mechanisms of perceptual priming , 1998, Current Opinion in Neurobiology.

[12]  Terrence J. Sejnowski,et al.  Unsupervised Learning , 2018, Encyclopedia of GIS.

[13]  Geoffrey E. Hinton,et al.  Unsupervised learning : foundations of neural computation , 1999 .

[14]  Geoffrey E. Hinton,et al.  Developing Population Codes by Minimizing Description Length , 1993, NIPS.

[15]  K J Holyoak,et al.  Distributional expectations and the induction of category structure. , 1986, Journal of experimental psychology. Learning, memory, and cognition.

[16]  Nathan Intrator,et al.  A Productive, Systematic Framework for the Representation of Visual Structure , 2000, NIPS.

[17]  Dean Pomerleau,et al.  Input Reconstruction Reliability Estimation , 1992, NIPS.

[18]  G. Berns,et al.  Brain regions responsive to novelty in the absence of awareness. , 1997, Science.