Elias B. Issa

This paper shows that training DNNs on face geometry and emotion improves generalization to new face tasks with fewer samples than standard models.

The paper shows that humans and marmosets perform similarly in shape-based face recognition tasks, indicating shared high-level visual processing capabilities.

This study shows that factorization of visual scene information is a common strategy in primate brains and deep neural networks, improving object identity decoding.

This study compares visual recognition abilities across species, finding that brain organization, not size, correlates with high-level visual performance in marmosets.

Among dozens of ImageNet-trained networks scored against macaque V4/IT recordings and human behavior, DenseNet-169, CORnet-S and ResNet-101 rank highest, and brain similarity stops tracking accuracy above 70% top-1.

The paper shows that recurrent circuits in the primate ventral stream are critical for object recognition, as challenge images require longer processing and are poorly predicted by feedforward models.

The paper shows that neural dynamics in the ventral visual stream involve hierarchical error signals between cortical stages, captured by recurrent models rather than standard feedforward circuits.

This paper compares human, monkey, and deep neural network performance in object recognition at the image level, finding that current models fail to predict primate behavior.