지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학술저널
- 저자정보
- 저널정보
- 대한전자공학회 JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE Journal of Semiconductor Technology and Science Vol.18 No.5
- 발행연도
- 2018.10
- 수록면
- 568 - 577 (10page)
- DOI
- 10.5573/JSTS.2018.18.5.568
이용수
초록· 키워드
오류제보하기
Binarized Neural Networks (BNNs) use only binary synapses of +1 and -1, not allowing any intermediate weights between -1 and +1. Though the recognition rate of BNNs is lower than the conventional Deep Neural Networks (DNNs), BNNs have attracted many interests nowadays, because BNNs do not need the complicated multiplication such as DNNs. Binary memristor crossbars can be very suitable to realize BNN hardware. This is because, in memristor BNNs, simple binary operation can be performed in bitwise manner for all the columns in memristor crossbars, simultaneously. In this paper, single-column and double-column memristor BNNs are presented, respectively. In addition, ReLU and sigmoid activation function circuits are also proposed with CMOS circuits. The designed Memristor-CMOS hybrid circuits of BNNs have been tested for MNIST vectors. The memristor BNNs could recognize almost 90% MNIST digits when the memristance variation is as large as 25%. For variation tolerance, the memristor BNNs are compared with the multi-valued memristor neural networks such as 4-bit, 6-bit, etc, in this paper. As a result, it has been confirmed the memristor BNNs become more variation-tolerant than the multi-valued memristor NNs when the variation becomes larger than 22%. Comparing the single-column and doublecolumn BNNs in this paper indicates that the singlecolumn BNN can save power consumption and array area almost by half than the double-column. This is because the single-column has just half memristors than the double-column. And, we measured the single-column and double-column BNNs using the fabricated memristor array. In this measurement, both the double-column and single-column BNNs were observed to work well.
목차
Abstract
I. INTRODUCTION
I. MEMRISTOR BINARIZED NEURAL NETWORKS
II. MEMRISTOR CROSSBARS AND ACTIVATION FUNCTION CIRCUITS
III. SIMULATION AND EXPERIMENTAL RESULTS
IV. CONCLUSION
REFERENCES
참고문헌 (20)
참고문헌 신청
A. Krizhevsky, I. Sutskever, and G. E. Hinton, ImageNet Classification with Deep Convolutional Neural Networks, In Advances in Neural Information Processing Systems 25, pp. 1106?1114, 2012.
I. Hubara, M. Courbariaux, D. Soudry, R. El-Yaniv, and Y. Bengio, Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations, arXiv:1609.07061, 2016.
M. Courbariaux and Y. Bengio, Binarized Neural Networks: Training Neural Networks with Weights and Activations Constrained to +1 or ?1, arXiv:1602.02830, 2016.
M. Courbariaux, Y. Bengio, and J.-P. David, Binary Connect: Training Deep Neural Networks with binary weights during propagations, arXiv:1511.00363, 2016.
L. O. Chua, Memristor-the missing circuit element, IEEE Trans. Circuit Theory, vol. CT-18, no. 5, pp. 507?519, 1971.
함께 읽어보면 좋을 논문
논문 유사도에 따라 DBpia 가 추천하는 논문입니다. 함께 보면 좋을 연관 논문을 확인해보세요!
-
Memristor crossbar array for binarized neural networks
대한전자공학회 학술대회
2019 .06
-
Simplified Design Procedure for Reinforced Concrete Columns Based on Equivalent Column Concept
International Journal of Concrete Structures and Materials
2016 .09
-
Buckling resistance of axially loaded square concrete-filled double steel tubular columns
Steel and Composite Structures, An International Journal
2022 .06
-
Mechanical properties of reinforced-concrete rocking columns based on damage resistance
Structural Engineering and Mechanics, An Int＇l Journal
2021 .12
-
메모리 감소와 저전력을 위한 BNN 가속기의 구조 최적화
전자공학회논문지
2022 .08
-
Hysteretic behaviors and calculation model of steel reinforced recycled concrete filled circular steel tube columns
Structural Engineering and Mechanics, An Int＇l Journal
2022 .08
-
Stress-transfer in concrete encased and filled tube square columns employed in top-down construction
Steel and Composite Structures, An International Journal
2016 .01
-
Dynamic experimental study on single and double beam-column joints in steel traditional-style buildings
Structural engineering and mechanics : An international journal
2017 .01
-
Boundary Conditions and Fire Behavior of Concrete Filled Tubular Composite Columns
International journal of high-rise buildings
2018 .01
-
Study and design of assembled CFDST column-beam connections considering column wall failure
Steel and Composite Structures, An International Journal
2021 .01
-
수직하중에 대한 한옥 기둥부재의 초기설계를 위한 구조검토방안
대한건축학회 논문집 - 구조계
2015 .11
-
Dynamic experimental study on single and double beam-column joints in steel traditional-style buildings
Structural Engineering and Mechanics, An Int＇l Journal
2017 .01
-
Evaluation of monotonic and cyclic behaviour of geotextile encased stone columns
Structural Engineering and Mechanics, An Int＇l Journal
2018 .01
-
Study on mechanical behaviors of column foot joint in traditional timber structure
Structural engineering and mechanics : An international journal
2018 .01
-
정전 컬럼 제어를 위한 프로토콜 설계
한국정밀공학회 학술발표대회 논문집
2015 .12
-
[시니어 컬럼] 행복의 구성요소는 무엇인가?
건축
2016 .12
-
Simplified Effective Compressive Strengths of Columns with Intervening Floor Slabs
International Journal of Concrete Structures and Materials
2020 .09
-
초소형 전자빔을 이용한 멀티 전자빔 운영 시스템 설계
한국레이저가공학회지
2015 .01
이 논문의 저자 정보
이 논문과 함께 이용한 논문
New Memristor-Based Crossbar Array Architecture with 50-% Area Reduction and 48-% Power Saving for Matrix-Vector Multiplication of Analog Neuromorphic Computing
Son Ngoc T
,
uong
,
Kyeong-Sik Min
JOURNAL OF SEMICONDUCTOR TECHNOLOGY AND SCIENCE
2014
.06
하드웨어 기반 스파이킹 뉴럴 네트워크를 이용한 강화 학습 검증에 대한 연구
오민혜
,
장지수
,
박병국
대한전자공학회 학술대회
2018
.06
멤리스터 소자를 활용한 뉴로모픽 동향연구
유호영
,
김영민
전자공학회지
2018
.07
최근 본 자료
댓글(0)
0
UCI(KEPA) : I410-ECN-0101-2019-569-000047808