
배민호
Minho Bae
minkkang1 at naver.com
After graduation
한국전자통신연구원 (ETRI)
Publications
2022
여상호,; 배민호,; 정민중,; 권오경,; 오상윤,
Crossover-SGD: A gossip-based communication in distributed deep learning for alleviating large mini-batch problem and enhancing scalability Journal Article
In: Concurrency and Computation: Practice and Experience, 2022.
@article{여상호2022Crossover-SGD,
title = {Crossover-SGD: A gossip-based communication in distributed deep learning for alleviating large mini-batch problem and enhancing scalability},
author = {여상호 and 배민호 and 정민중 and 권오경 and 오상윤},
url = {https://arxiv.org/abs/2012.15198},
doi = {10.48550/arXiv.2012.15198},
year = {2022},
date = {2022-11-01},
urldate = {2022-11-01},
journal = {Concurrency and Computation: Practice and Experience},
abstract = { Distributed deep learning is an effective way to reduce the training time of deep learning for large datasets as well as complex models. However, the limited scalability caused by network overheads makes it difficult to synchronize the parameters of all workers. To resolve this problem, gossip-based methods that demonstrates stable scalability regardless of the number of workers have been proposed. However, to use gossip-based methods in general cases, the validation accuracy for a large mini-batch needs to be verified. To verify this, we first empirically study the characteristics of gossip methods in a large mini-batch problem and observe that the gossip methods preserve higher validation accuracy than AllReduce-SGD(Stochastic Gradient Descent) when the number of batch sizes is increased and the number of workers is fixed. However, the delayed parameter propagation of the gossip-based models decreases validation accuracy in large node scales. To cope with this problem, we propose Crossover-SGD that alleviates the delay propagation of weight parameters via segment-wise communication and load balancing random network topology. We also adapt hierarchical communication to limit the number of workers in gossip-based communication methods. To validate the effectiveness of our proposed method, we conduct empirical experiments and observe that our Crossover-SGD shows higher node scalability than SGP(Stochastic Gradient Push). },
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2020
민경준,; 배민호,; 박규동,; 오상윤,
혼합형 출판-구독 메시징 시스템 구조 제안 및 연동 기법 Journal Article
In: 한국통신학회논문지, vol. 45, no. 4, pp. 748–755, 2020.
@article{민경준2020혼합형,
title = {혼합형 출판-구독 메시징 시스템 구조 제안 및 연동 기법},
author = {민경준 and 배민호 and 박규동 and 오상윤},
url = {http://www.dbpia.co.kr/journal/articleDetail?nodeId=NODE09325603},
year = {2020},
date = {2020-01-01},
journal = {한국통신학회논문지},
volume = {45},
number = {4},
pages = {748--755},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2019
배민호,; 민경준,; 박규동,; 전호철,; 오상윤,
군 통신환경을 효과적으로 활용하는 하이브리드 메시징 시스템 구조와 XML Element 기반의 필터링 기법 Journal Article
In: 정보과학회논문지, vol. 46, no. 10, pp. 1078–1087, 2019.
@article{배민호2019군,
title = {군 통신환경을 효과적으로 활용하는 하이브리드 메시징 시스템 구조와 XML Element 기반의 필터링 기법},
author = {배민호 and 민경준 and 박규동 and 전호철 and 오상윤},
year = {2019},
date = {2019-01-01},
urldate = {2019-01-01},
journal = {정보과학회논문지},
volume = {46},
number = {10},
pages = {1078--1087},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
2017
김동훈,; 배민호,; 오상윤,; 정민중,
성능평가를 통한 컴퓨터 비전 프로세스의 임베디드 GPU 환경 적용 분석 Conference
2017년 한국컴퓨터종합학술대회, 2017.
@conference{김동훈2017성능평가를,
title = {성능평가를 통한 컴퓨터 비전 프로세스의 임베디드 GPU 환경 적용 분석},
author = {김동훈 and 배민호 and 오상윤 and 정민중},
year = {2017},
date = {2017-01-01},
urldate = {2017-01-01},
booktitle = {2017년 한국컴퓨터종합학술대회},
journal = {한국정보과학회 학술발표논문집},
pages = {1477--1479},
keywords = {},
pubstate = {published},
tppubtype = {conference}
}
2014
엄준호,; 배민호,; 오상윤,
클라우드 컴퓨팅 기반 해군 함정 상태정보 공유 아키텍처 제안 Conference
한국군사기술과학회 종합학술대회, 2014.
@conference{엄준호2014클라우드,
title = {클라우드 컴퓨팅 기반 해군 함정 상태정보 공유 아키텍처 제안},
author = {엄준호 and 배민호 and 오상윤},
year = {2014},
date = {2014-01-01},
booktitle = {한국군사기술과학회 종합학술대회},
keywords = {},
pubstate = {published},
tppubtype = {conference}
}