CNNs reveal the computational implausibility of the expertise hypothesis
Nancy Kanwisher,
Pranjul Gupta,
Katharina Dobs
Affiliations
Nancy Kanwisher
Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA; McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
Pranjul Gupta
Department of Psychology, Justus-Liebig University Giessen, 35394 Giessen, Germany
Katharina Dobs
Department of Psychology, Justus-Liebig University Giessen, 35394 Giessen, Germany; Center for Mind, Brain and Behavior (CMBB), University of Marburg and Justus-Liebig University, 35032 Marburg, Germany; Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA; McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA; Corresponding author
Summary: Face perception has long served as a classic example of domain specificity of mind and brain. But an alternative “expertise” hypothesis holds that putatively face-specific mechanisms are actually domain-general, and can be recruited for the perception of other objects of expertise (e.g., cars for car experts). Here, we demonstrate the computational implausibility of this hypothesis: Neural network models optimized for generic object categorization provide a better foundation for expert fine-grained discrimination than do models optimized for face recognition.
