Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study investigates the relationship between facial attractiveness and processing fluency by training variational autoencoders (VAEs) on multiple face datasets without direct attractiveness supervision. The findings reveal that human attractiveness ratings correlate significantly with the VAE evidence lower bound (ELBO) in rate-distortion space, suggesting that attractive faces are perceived more easily due to their prototypical features. Additionally, the research demonstrates that the attractiveness direction in the latent space is robust across different model initializations and datasets, linking aesthetic appreciation to generative modeling frameworks.
Human ratings of facial attractiveness align closely with a latent representation defined by variational autoencoders, revealing that beauty is fundamentally tied to processing fluency.
Facial attractiveness has been linked to statistical regularities such as symmetry and averageness, suggesting that beauty may depend on the ease with which a face is perceived. We empirically test this hypothesis by training variational autoencoders on four face datasets without attractiveness supervision and evaluating their representations on the 597 faces from the Chicago Face Database. Across models, human attractiveness ratings closely aligns with the direction defined by the VAE evidence lower bound (ELBO) in rate-distortion space. Independently learned latent spaces contain an attractiveness direction that transfers strongly across random initializations and training data. We also find that attractive faces are more prototypical in both shape and latent space. Our results connect classic accounts of aesthetics with learned generative models and provide empirical support for a variational interpretation of the processing fluency theory of aesthetic pleasure.