id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_5tepnlhakbaobhkcun7emmiwwa	Julien P. Renoult	Beauty is in the efficient coding of the beholder	2016	8	.pdf	application/pdf	5477	919	64	sexual selection, sparse coding, sensory bias, statistics of the image, leaving the higher-order redundancies that are analysed through sparse coding. To study the sparseness of faces, we first trained an artificial neural network to reconstruct whitened images of natural scenes with a sparse coding algorithm, a step known as dictionary learning in with attractiveness as a response variable, facial symmetry, skin roughness, age of women and sparseness predicted that stimuli coded sparsely by the perceptual system should be viewed as attractive [34]. We evidenced that faces coded sparsely by the primary visual cortex V1 are more attractive. The positive correlation between face attractiveness and sparse coding in our model of V1 simple cells thus indicates that attraction for faces is, at least in part, a efficient coding bias and of preferences that evolved to assess mate quality in sexual selection. learning sparse codes for natural images.	./cache/work_5tepnlhakbaobhkcun7emmiwwa.pdf	./txt/work_5tepnlhakbaobhkcun7emmiwwa.txt