Abstract:Generations of scientists have pursued the goal of defining beauty. While early scientists initially focused on objective criteria of beauty ('feature-based aesthetics'), philosophers and artists alike have since proposed that beauty arises from the interaction between the object and the individual who perceives it. The aesthetic theory of fluency formalizes this idea of interaction by proposing that beauty is determined by the efficiency of information processing in the perceiver's brain ('processing-based aesthetics'), and that efficient processing induces a positive aesthetic experience. The theory is supported by numerous psychological results, however, to date there is no quantitative predictive model to test it on a large scale. In this work, we propose to leverage the capacity of deep convolutional neural networks (DCNN) to model the processing of information in the brain by studying the link between beauty and neuronal sparsity, a measure of information processing efficiency. Whether analyzing pictures of faces, figurative or abstract art paintings, neuronal sparsity explains up to 28% of variance in beauty scores, and up to 47% when combined with a feature-based metric. However, we also found that sparsity is either positively or negatively correlated with beauty across the multiple layers of the DCNN. Our quantitative model stresses the importance of considering how information is processed, in addition to the content of that information, when predicting beauty, but also suggests an unexpectedly complex relationship between fluency and beauty. Developing good predictive models of beauty requires understanding what happens in the brain when we find a person or an artwork beautiful. Recent theories in psychology emphasize the importance of considering how the brain processes features, in addition to the features themselves. Features that are efficiently processed by the brain, such as symmetry, fractality, or naturalness are generally perceived as visually attractive. In this study, we leveraged the capacity of artificial intelligence to model information processing in the human brain, to evaluate how the beauty of human faces and artistic paintings can be predicted from the efficiency of the neural code. Our results show that the efficiency of information processing can explain approximately one-third of the perception of beauty and emphasize the importance of considering how information is processed when investigating beauty. Additionally, our use of artificial intelligence demonstrates the potential of this technology to help better understand complex human behaviors.

Biologically-inspired characterization of sparseness in natural images

Differential response of the retinal neural code with respect to the sparseness of natural images

Efficient Coding for Natural Images Based on the Sparseness of Neural Coding in V1 Across the Stimuli

A Sparse Coding Model with Synaptically Local Plasticity and Spiking Neurons Can Account for the Diverse Shapes of V1 Simple Cell Receptive Fields

Kernel Heterogeneity Improves Sparseness of Natural Images Representations

Sparsity in an artificial neural network predicts beauty: Towards a model of processing-based aesthetics

Sparsity and Compressed Coding in Sensory Systems

Beyond $\ell_1$ sparse coding in V1

Emergence of simple-cell receptive field properties by learning a sparse code for natural images

New Evidences for Sparse Coding Strategy Employed in Visual Neurons: from the Image Processing and Nonlinear Approximation Viewpoint.

A Lateral Inhibitory Spiking Neural Network for Sparse Representation in Visual Cortex.

Sparse Representation and Synaptic Adaptation of the Visual Sensory System

Emergence of Sparse Coding, Balance and Decorrelation from a Biologically-Grounded Spiking Neural Network Model of Learning in the Primary Visual Cortex

Role of homeostasis in learning sparse representations

Sparse Modeling for Image and Vision Processing

A Unified Theory of Response Sparsity and Variability for Energy-Efficient Neural Coding

On Sparse Representations of Color Images

Efficient Representation of Natural Image Patches

Adaptation to visual sparsity enhances responses to infrequent stimuli

Sparsey™: event recognition via deep hierarchical sparse distributed codes

A Model of Attention-Guided Visual Sparse Coding