A brain-inspired network architecture for cost-efficient object recognition in shallow hierarchical neural networks

Youngjin Park, Seungdae Baek, Se-Bum Paik

Abstract: Highlights•Long-range connections (LRCs) enable recognition of various objects under constraints.•LRCs added to shallow networks enhance performance comparable to deeper networks.•LRCs can emerge spontaneously while balancing performance and wiring cost.•LRCs reveal a biological strategy for a cost-efficient neural network architecture.