About this document
Evolvable Block-Based Neural Network Design by chagoppss is a document available to read on EtoBox.
This paper discusses the hardware implementation of evolvable block-based neural networks (BbNNs) that utilize a cost-efficient sigmoid-like activation function, focusing on optimizing structure for reconfugurable digital hardware like FPGAs. The proposed design features neuron blocks with a single multiplier and presents a novel method for approximating the hyperbolic tangent function, enhancing performance while minimizing resource utilization. The BbNN system is designed as a system-on-chip (SoC) and has
- Author
- chagoppss
- Language
- EN