.png){kind=icon}
返回
Quality Driven Systematic Approximation for Binary-Weight Neural Network Deployment
.png){kind=icon}
PRE
AI
DOI：10.1109/TCSI.2022.3164170.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Neural networks (NNs) with large scales of artificial neurons are increasingly used in recognition and classification tasks. In power-constrained scenarios, the tradeoff between performance and hardware consumptions must be carefully evaluated before silicon tape-out. In this paper, we proposed a systematic approach to design ultra-low power NN system. This work is motivated by the facts that NNs are resilient to approximation in many of the computations and NNs are outputting statistical tensors which are acceptable to less-than-perfect results. We resort to the front-back end approach with a twofold aim: (1) a fast and accurate design approach is proposed by estimating the computing quality of low-power approximate adder arrays, and it is adopted to evaluate the neural network system; (2) a quality configurable engine with different approximation degrees while processing NNs is implemented. The proposed work is demonstrated with a comprehensive keyword spotting (KWS) system as an ultra-low power NN engine. The experimental environment is setup with ten keywords from the google speech command dataset (GSCD) using an industrial 22-nm ultra-low-leakage (ULL) process. Comparing to the state-of-the-art KWS processors, the proposed approximate NN engine can demonstrate over 60% improvement in power efficiency and 1.1x area efficiency while achieving similar recognition accuracy.
Keyword:
Adders
Artificial neural networks
Computational modeling
Approximate computing
Estimation
Integrated circuit modeling
Analytical models
Approximate computing
neural network
BNN
low power accelerator
deployment
期刊
IF:
5.2
论文数:
9.8K
被引数:
2.2W
机构
引用论文
分享收藏A Hierarchical Cardiac Rhythm Classification Methodology Based on Electrocardiogram Fiducial Points一种基于心电图基准点的分层心律分类方法
OAAI
OAAI分享收藏A Retrospective and Prospective View of Approximate Computing [Point of View}.png)
PROCEEDINGS OF THE IEEEOAAI
PROCEEDINGS OF THE IEEEIF25.9
OAAI分享收藏A Theoretical Framework for Quality Estimation and Optimization of DSP Applications Using Low-Power Approximate Adders使用低功耗近似加法器的DSP应用质量估计和优化的理论框架
PREAI
分享收藏
查看更多内容