Design Space Exploration of Neural Network Activation Function Circuits

TL;DR

This paper presents a highly efficient combinational circuit approach for implementing activation functions like SELU and tanh in neural network accelerators, achieving significant area and speed improvements with minimal accuracy loss.

Contribution

It introduces a novel combinational circuit design methodology for activation functions, optimizing hardware efficiency and demonstrating negligible impact on neural network accuracy.

Findings

Circuit designs for tanh and SELU save significant area compared to LUT-based implementations.

The proposed circuits operate at high frequencies (over 4 GHz).

Neural networks show insensitivity to activation function precision.

Abstract

The widespread application of artificial neural networks has prompted researchers to experiment with FPGA and customized ASIC designs to speed up their computation. These implementation efforts have generally focused on weight multiplication and signal summation operations, and less on activation functions used in these applications. Yet, efficient hardware implementations of nonlinear activation functions like Exponential Linear Units (ELU), Scaled Exponential Linear Units (SELU), and Hyperbolic Tangent (tanh), are central to designing effective neural network accelerators, since these functions require lots of resources. In this paper, we explore efficient hardware implementations of activation functions using purely combinational circuits, with a focus on two widely used nonlinear activation functions, i.e., SELU and tanh. Our experiments demonstrate that neural networks are generally insensitive to the precision of the activation function. The results also prove that the proposed combinational circuit-based approach is very efficient in terms of speed and area, with negligible accuracy loss on the MNIST, CIFAR-10 and IMAGENET benchmarks. Synopsys Design Compiler synthesis results show that circuit designs for tanh and SELU can save between 3.13-7.69 and 4.45-8:45 area compared to the LUT/memory-based implementations, and can operate at 5.14GHz and 4.52GHz using the 28nm SVT library, respectively. The implementation is available at: https://github.com/ThomasMrY/ActivationFunctionDemo.