Journal of the Institute of Electronics Engineers of Korea CI (전자공학회논문지CI)

The Institute of Electronics and Information Engineers (대한전자공학회)
권호 표지

Expansible & Reconfigurable Neuro Informatics Engine : ERNIE

대규모 확장이 가능한 범용 신경망 연산기 : ERNIE

  • 김영주 (인하대학교 정보통신공학과) ;
  • 동성수 (인하대학교 정보통신공학과) ;
  • 이종호 (용인송담대학 디지털전자정보과)
  • Published : 2003.11.01
Download PDF
⟨ Previous Next ⟩

Abstract

Difficult problems In implementing digital neural network hardware are the extension of synapses and the programmability for relocating neurons. In this paper, the structure of a new hardware is proposed for solving these problems. Our structure based on traditional SIMD can be dynamically and easily reconfigured connections of network without synthesizing and mapping original design for each use. Using additional modular processing unit the numbers of neurons find synapses increase. To show the extensibility of our structure, various models of neural networks : multi-layer perceptrons and Kohonen network are formed and tested. The performance comparison with software simulation shows its superiority in the aspects of performance and flexibility.

범용 신경망 연산기를 디지털 회로로 구현함에 있어 가장 까다로운 문제들 중 하나는 시냅스의 확장과 해당 네트워크에 맞게 뉴런들을 재배치하는 재구성 문제일 것이다. 본 논문에서는 이러한 문제들을 해결하기 위한 새로운 하드웨어 구조를 제안한다. 제안된 구조는 시냅스의 확장과 네트워크 구조의 변경을 위해 오리지날 디자인의 변경이 필요치 않으며, 모듈러 프로세싱 유니트의 확장을 통한 뉴런의 개수 및 레이어의 확장이 가능하다. 이 구조의 범용성 및 확장성에 대한 검증을 위해 다양한 종류의 다층 퍼셉트론 및 코호넨 네트워크를 구성하여 HDL 시뮬레이터를 통한 결과와 C 언어로 작성된 소프트웨어 시뮬레이터 결과를 비교하였으며 그 결과 성능이 거의 일치함을 확인하였다.

Keywords

References

  1. Robert J. Schalkoff, 'Artificial neural networks', McGraw-Hill, 1997
  2. Yingang Wang, Zuocheng Ma, Huaxiang Lu, Shoujue Wang, 'Discussion on the methodology of neural network hardware design and implemetnation', Solid-State and Integrated-Circuit Technology, 2001. Proceedings. 6th Intermational Conferences on, vol.1 pp. 113-116, 2001 https://doi.org/10.1109/ICSICT.2001.981435
  3. Miroslav Skrbek, 'Fast neural network implementation', Neural Network World, pp. 357-391, 1999
  4. Tams Szab, Lrinc Antoni, Gbor Horvth, Bla Fehr, 'A full-parallel digital implementation for pre-trained NNs', Neural Network, 2000. IJCNN 200, Proceedings of the IEEE-INNS-ENNS Intemational Joint Conference on 2000, vol.2, pp. 49-54, 2000 https://doi.org/10.1109/IJCNN.2000.857873
  5. Masahiro Murakawa, Shuji Yoshizawa, Isamu Kajitani, Xin Yao, Nobuki Kajihara, Masaya Iwata and Tetsuya Higuchi, 'The GRD Chip : Genetic Reconfiguration of DSPs for Neural Network Processing', IEEE Transaction on computers, vol.48, no.6, June 1999 https://doi.org/10.1109/12.773799
  6. Bernard Girau, 'Digital hardware implementation of 2D complatibel neural netwroks', Neural Network, 2000. IJCNN 2000. Proceedings of the IEEE-INNS-ENNS International Joint Conference on 2000, Volume: 3, pp. 506-511, 2000 https://doi.org/10.1109/IJCNN.2000.861359
  7. B. Pino, F.J. Pelayo, J. Ortega and A. Prieto, 'Design and Evaluation of a Reconfigurable Digital Architecture for Self-Organizing Maps', Microelectronics for Neural, Fuzzy and BioI-Inspired Systmes, MciroNeuro '99. Proceedings of the Seventh International Conferences on 1999, pp. 395-402, 1999 https://doi.org/10.1109/MN.1999.758892
  8. S. Vitabile, A. Gentile, G.B. Dammone, F. Sorbello, 'Multi-layer perceptron mapping on a SIMD architecture', Neural Networks of Signal Processing, Proceedings of the 2002 12th IEEE Workshop, pp. 667-675, 2002 https://doi.org/10.1109/NNSP.2002.1030078
  9. Simon Haykin, 'Neural networks a comprehensive foundation', Prentice hall, pp. 135, pp.448, 1999
  10. Nikola B. Serbedzija, 'Simulating Artificial Neural Networks on Parallel Architectures', Computer, vol. 29, Issue: 3, March 1996, pp. 56-63 https://doi.org/10.1109/2.485893
  11. O. L. Mangasarian and W. H. Wolberg : 'Cancer diagnosis via linear programming', SIAM News, vol.3, Number 5, September 1990