Changement d'a priori : une théorie bayésienne de la perception et de l'apprentissage dans l'autisme
Résumé
More than 10 years ago, David Burr and I proposed the use of computational modelling to identify the mechanisms that underlie autistic sensation and perception. Specifically, we suggested that attenuated Bayesian priors could result in autistic people’s unique perceptual experience: the tendency to perceive the world “as it really is” immediately rather than imbued by prior experience. In this talk, I review the academic response to this proposal over the ensuing decade and trace the way in which it has reshaped the scholarly community’s understanding of autism.
