Efficient Soft-Error Detection for Low-precision Deep Learning Recommendation Models

Author:

Li Sihuan1,Huang Jianyu1,Tang Ping Tak Peter1,Khudia Daya2,Park Jongsoo1,Dixit Harish Dattatraya1,Chen Zizhong3

Affiliation:

1. Meta Platforms, Inc.

2. MosaicML, Inc.

3. University of California,Riverside,CA,USA

Funder

U.S. Department of Energy

Office of Science

Advanced Scientific Computing Research

Publisher

IEEE

Reference32 articles.

1. New-Sum

2. Online-ABFT

3. Algorithm-Based Fault Tolerance for Matrix Operations

4. Making convolutions resilient via algorithm-based error detection techniques;hari,2020

5. Post-training 4-bit quantization on embedding tables;guan,2019

Cited by 3 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Brief Industry Paper: Evaluating Robustness of Deep Learning-Based Recommendation Systems Against Hardware Errors: A Case Study;2023 IEEE Real-Time Systems Symposium (RTSS);2023-12-05

2. Soft Error Reliability Analysis of Vision Transformers;IEEE Transactions on Very Large Scale Integration (VLSI) Systems;2023-12

3. Exploring Winograd Convolution for Cost-Effective Neural Network Fault Tolerance;IEEE Transactions on Very Large Scale Integration (VLSI) Systems;2023-11

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3