Answers are grounded in this paper and cite their sources.
Using the replica method the authors investigate learning from correlated data in a perceptron. Both types of association-'spatial'-within each example, and 'semantic'-among the different ones-are considered for both Boolean and continuous weights and output functions. 'Spatial' correlations of data may significantly improve the learning properties, whereas the 'semantic' ones do not practically influence them.
No takes yet. Share an insight, caveat, or question.
Tarkowski et al. (1993) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: