Search Papers | Poster Sessions | All Posters
Poster B73 in Poster Session B - Thursday, August 8, 2024, 1:30 – 3:30 pm, Johnson Ice Rink
Recurrent models optimized for face recognition exhibit representational dynamics resembling the primate brain
Hossein Adeli1 (), Nikolaus Kriegeskorte1; 1Columbia University
Understanding the dynamics of neural representations is crucial for elucidating the mechanisms of visual recognition in the primate brain. Here we investigate the representational dynamics of recurrent convolutional neural networks (RCNNs) optimized for face-identification and object-recognition tasks. Using representational similarity analysis (RSA), we observed that only models that were trained for face identification showed a late-emerging prominent distinction of identities as seen in the monkey face patch AM. Interestingly, early model responses (to a diverse set of images including human faces, monkey faces, and non-face objects) strongly separated the objects from faces. Our results also show that models that were trained simultaneously on both face identification and object recognition were more likely to show the signature of mirror symmetric viewpoint tuning in their intermediate representations as has been reported for monkey face patch AL. These findings suggest that the dynamics of face recognition that emerges in a hierarchical recurrent neural network prioritizes category-level recognition at early stages, triggering category-specific computations that enable individual-level recognition.
Keywords: Face perception Recurrent processing Deep Neural Networks