DOI QR코드

DOI QR Code

Improvement of Properties of the Fuzzy ART with the Variable Weighed Average Learning

가변 가중 평균 학습을 적용한 퍼지 ART 신경망의 성능 향상

  • Received : 2016.10.12
  • Accepted : 2017.01.09
  • Published : 2017.02.28

Abstract

In this paper, we propose a variable weighted average (VWA) learning method in order to improve the performance of the fuzzy ART neural network that has been developed by Grossberg. In a conventional method, the Fast Commit Slow Recode (FCSR), when an input pattern falls in a category, the representative pattern of the category is updated at a fixed learning rate regardless of the degree of similarity of the input pattern. To resolve this issue, a variable learning method proposes reflecting the distance between the input pattern and the representative pattern to reduce the FCSR's category proliferation issue and improve the pattern recognition rate. However, these methods still suffer from the category proliferation issue and limited pattern recognition rate due to inevitable excessive learning created by use of fuzzy AND. The proposed method applies a weighted average learning scheme that reflects the distance between the input pattern and the representative pattern when updating the representative pattern of a category suppressing excessive learning for a representative pattern. Our simulation results show that the newly proposed variable weighted average learning method (VWA) mitigates the category proliferation problem of a fuzzy ART neural network by suppressing excessive learning of a representative pattern in a noisy environment and significantly improves the pattern recognition rates.

본 논문은 그로스버그(Grossberg)에 의해 개발된 퍼지 ART 신경 회로망의 성능을 향상시키기 위하여 가변가중 평균(VWA) 학습 방법을 제안한다. 기존의 방법인 고속수용저속부호화(FCSR)는 입력패턴이 임의의 카테고리 내에 포함될 때 카테고리를 대표하는 대표패턴의 갱신이 입력패턴과의 거리(유사성)와 관계없이 고정 학습률로 갱신되고, 또한 이를 개선한 가변학습(VL)은 대표패턴과 입력패턴 사이의 거리를 대표패턴의 갱신에 반영하여 카테고리 증식 문제와 패턴 인식률을 개선한다. 그러나 두 방법 모두 학습 시 퍼지 AND에 의한 과도한 학습이 필수적으로 발생하여 카테고리 증식 문제와 패턴 인식 향상에 한계를 갖는다. 제안된 방법은 카테고리를 대표하는 대표패턴의 갱신 시 대표패턴과 입력패턴 사이의 거리를 반영한 가중평균 학습을 적용하여 대표패턴의 과도한 학습을 억제한다. 시뮬레이션 결과 기존의 학습 방법인 고속수용저속부호화(FCSR)와 가변학습(VL) 보다 제안된 가변가중평균(VWA) 학습 방법이 잡음 환경에서 대표패턴의 과도한 학습을 억제하여 퍼지 ART 신경 회로망의 카테고리 증식문제를 완화하고 패턴 인식률을 향상시키는 것을 보여준다.

Keywords

References

  1. G. A. Carpenter and S. Grossberg, "A massively parallel architecture for a selforganizing neural pattern recognition machine," Comput. Vision, Graphics, Image Process., vol. 37, no. 1, pp. 54-115, Jan. 1987. https://doi.org/10.1016/S0734-189X(87)80014-2
  2. G. A. Carpenter and S. Grossberg, "Fuzzy ART: Fast stable learning and categorization of analog patterns by an adaptive resonance system," Neural Networks, vol. 4, no. 6, pp. 759-771, Dec. 1991. https://doi.org/10.1016/0893-6080(91)90056-B
  3. G. A. Carpenter, S. Grossberg, N. Markuzon, J. Reynolds, and D. Rosen, "Fuzzy ARTMAP: A neural network architecture for incremental supervised learning of multidimensional maps," IEEE Trans. Neural Networks, vol. 3, pp. 698-713, Sept. 1992. https://doi.org/10.1109/72.159059
  4. G. A. Carpenter and M. N. Gjaja, "Fuzzy ART choice functions," in Proc. WCNN, pp. 713-722, San Diego, USA, Jun. 1994.
  5. C. J. Lee, S. Y. Lee, and C. W. Lee, "Improvement of noise tolerance in Fuzzy ART using a weighted sum and a fuzzy and operation," IEICE Trans. Fundamentals Electron., Commun., Comput. Sci., vol. E78-A, no. 10, pp. 1432-1434, Oct. 1995.
  6. C. J. Lee, S. Y. Lee, and C. W. Lee, "A study on the new learning method to improve noise tolerance in fuzzy ART," J. IEEK, vol. 32-B, no. 10, pp. 112-117, Oct. 1995.
  7. C. J. Lee, C. G. Yoon, and C. W. Lee, "A new learning method to improve category proliferation problem in fuzzy ART," in Proc. IEEE Int. Conf. Neural Network, vol. 3, pp. 1393-1396, Perth, Australia, Nov. 1995.
  8. J. S. Lee, C. J. Lee, and C. W. Lee "Improvement of recognition performance for fuzzy ARTMAP using average learning and slow learning," IEICE Trans. Fundamentals Electron., Commun., Comput. Sci., vol. E81-A, no. 3, pp. 514-516, Mar. 1998.
  9. J. Lee, "Fuzzy-ARTMAP based multi-user detection," J. KICS, vol. 37A no. 3. pp. 172- 178, Mar. 2012.
  10. C. J. Lee, B. Son, and H. S. Hong "Improvement of pattern recognition capacity of the fuzzy ART with the variable learning," J. KICS, vol. 38B, no. 12. pp. 954-961, Nov. 2013. https://doi.org/10.7840/kics.2013.38B.12.954