Stochastic Multivariate Universal-Radix Finite-State Machine: a Theoretically and Practically Elegant Nonlinear Function Approximator

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Feng, Xincheng, Shen, Guodong, Hu, Jianhao, Li, Meng, Wong, Ngai
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916235070930944
author Feng, Xincheng
Shen, Guodong
Hu, Jianhao
Li, Meng
Wong, Ngai
author_facet Feng, Xincheng
Shen, Guodong
Hu, Jianhao
Li, Meng
Wong, Ngai
contents Nonlinearities are crucial for capturing complex input-output relationships especially in deep neural networks. However, nonlinear functions often incur various hardware and compute overheads. Meanwhile, stochastic computing (SC) has emerged as a promising approach to tackle this challenge by trading output precision for hardware simplicity. To this end, this paper proposes a first-of-its-kind stochastic multivariate universal-radix finite-state machine (SMURF) that harnesses SC for hardware-simplistic multivariate nonlinear function generation at high accuracy. We present the finite-state machine (FSM) architecture for SMURF, as well as analytical derivations of sampling gate coefficients for accurately approximating generic nonlinear functions. Experiments demonstrate the superiority of SMURF, requiring only 16.07% area and 14.45% power consumption of Taylor-series approximation, and merely 2.22% area of look-up table (LUT) schemes.
format Preprint
id arxiv_https___arxiv_org_abs_2405_02356
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Stochastic Multivariate Universal-Radix Finite-State Machine: a Theoretically and Practically Elegant Nonlinear Function Approximator
Feng, Xincheng
Shen, Guodong
Hu, Jianhao
Li, Meng
Wong, Ngai
Machine Learning
Artificial Intelligence
Nonlinearities are crucial for capturing complex input-output relationships especially in deep neural networks. However, nonlinear functions often incur various hardware and compute overheads. Meanwhile, stochastic computing (SC) has emerged as a promising approach to tackle this challenge by trading output precision for hardware simplicity. To this end, this paper proposes a first-of-its-kind stochastic multivariate universal-radix finite-state machine (SMURF) that harnesses SC for hardware-simplistic multivariate nonlinear function generation at high accuracy. We present the finite-state machine (FSM) architecture for SMURF, as well as analytical derivations of sampling gate coefficients for accurately approximating generic nonlinear functions. Experiments demonstrate the superiority of SMURF, requiring only 16.07% area and 14.45% power consumption of Taylor-series approximation, and merely 2.22% area of look-up table (LUT) schemes.
title Stochastic Multivariate Universal-Radix Finite-State Machine: a Theoretically and Practically Elegant Nonlinear Function Approximator
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2405.02356