▲ to top

Uehara Laboratory

Kiyohiko Uehara, Dr. Eng., Associate Professor

Ibaraki University, Japan
Research on Computational Intelligence

 
α-GEM Family, invented by Dr. K. Uehara:
New approaches for fuzzy inference and fuzzy rule learning
toward their use in artificial intelligence

 
The α-GEM family consists of the methods listed in the following table at present.
All of these methods were invented by Dr. K. Uehara.
Invented Method   Meaning of the Method Name Year: 1st Proposal Main References
(1) α-RITHM α-level-set and arithmetic-mean-based inference 1995 [4]
(2) α-GEM α-level-set and generalized-mean-based inference 1997 [6]
(3) α-GEMII α-level-set and generalized-mean-based inference with the proof of two-sided symmetry of consequences 2009 [8-12]
(4) α-GEMS α-cut and generalized-mean-based inference in synergy with composition 2013 [17]
(5) α-GEMST α-level-set and generalized-mean-based inference in synergy with composition via linguistic-truth value control 2016 [19]
(6) α-GEMMI α-level-set and generalized-mean-based inference with multi-level interpolation 2011 [13]
(7) α-GEMMIET α-level-set and generalized-mean-based inference with multi-level interpolation extended in the number of points for interpolation 2013 [16]
(8) α-GEMILIE α-level-set and generalized-mean-based inference with infinite-level interpolation 2013 [15]
(9) α-GEMINAS α-level-set and generalized-mean-based inference with fuzzy rule interpolation at an infinite number of activating points 2015 [18]
(10) α-GEMINAST α-level-set and generalized-mean-based inference in synergy between α-GEMINAS and the composition via LTV control 2017 [21]
(11) α-GEMI-ES α-GEMINAS-based local-evolution toward slight linearity for global smoothness 2017 [23]
(12) α-GEMI-SING α-GEMII and α-GEMINAS-based local-evolution toward slight linearity for global smoothness by using singleton input–output rules 2018 [24]
(13) α-FUZZI-ES α-weight-based fuzzy-rule independent evaluations 2019 [26]
(14) α-FUZZI-ES Learning α-FUZZI-ES-based fuzzy-rule learning 2019 [26]
(15) α-FUZZI-EX α-weight-based fuzzy-rule independent evaluations extended for fuzzy inputs 2020 [25]
(16) α-FUZZI-EX Learning α-FUZZI-EX-based fuzzy-rule learning 2020 [25]
(17) α-GEMII-X Denoising α-GEMII-based denoising to unify fuzzy inference and preprocessing for fuzzy rule optimization 2021 [27]
 

Relations Between the Methods in α-GEM Family [22]

Family Tree of α-GEM Family
The α-GEM family consists of fuzzy inference methods for non-sparse rule base and sparse rule base [22]. In addition, it includes methods for noise reduction and fuzzy rule learning by which fuzzy rules can be efficiently optimized. α-GEMII has played a central role in the growth of the α-GEM family.

α-GEMII [8-12]: Fuzzy Inference Method

α-GEMII
α-GEMII can mathematically prove the convexity of consequences by the effective use of the generalized mean, while the fuzzy constraints of given facts are propagated to those of the deduced consequences. Moreover, it can mathematically prove the symmetricity of consequences under some conditions that are axiomatically derived from the viewpoint of fuzzy inference [8-12]. The operational process of the fuzziness-propagation control is not depicted in the figure for ease of visibility.

α-FUZZI-ES Learning [26]: Fast Fuzzy-Rule Optimization

α-FUZZI-ES Learning Title α-FUZZI-ES Learning
α-FUZZI-ES learning makes it possible to optimize fuzzy rules independently of each other. This property reduces the dimensionality of the search space in finding the optimum fuzzy rules. Thereby, α-FUZZI-ES learning can attain fast convergence in fuzzy rule optimization. Moreover, α-FUZZI-ES learning can be efficiently performed with hardware in parallel to optimize fuzzy rules independently of each other. α-FUZZI-ES learning is especially effective when evaluation functions are not differentiable and derivative-based optimization methods cannot be applied to fuzzy rule learning; α-FUZZI-ES learning is superior in fast convergence, especially with the derivative-free optimization of fuzzy rules [26].

Demonstration of α-FUZZI-ES Learning: Application to Interval Prediction [26]

The animation below shows the transitional changes of prediction intervals optimized by using α-FUZZI-ES learning; Each fuzzy rule is tuned by an immune algorithm in the scheme of α-FUZZI-ES learning.
Animation Instruction 03

α-GEMI-ES [23]: Noise Reduction for Fuzzy Rule Optimization

α-GEMI-ES
α-GEMI-ES
In the initial stage, fuzzy rules for α-GEMI-ES are set by directly using learning data. Subsequently, α-GEMI-ES iteratively performs α-GEMINAS and reduces noise in each iteration. α-GEMI-ES provides a unified platform for fuzzy inference and fuzzy rule learning with noise-corrupted data. It is expected to prevent fuzzy rules from overfitting to noise in learning data, especially when only a small amount of learning data is available for fuzzy rule optimization in comparison with the number of fuzzy rules [23].

Demonstration of α-GEMI-ES in Reducing Noise [23]

The animation below shows the transitional changes of noise reduction by α-GEMI-ES.
Animation Instruction 03

Main Papers

 More papers are listed here.
  1. K. Uehara, E. Taguchi, T. Watahiki, and T. Miyata, “A Data Converter for Analog/Fuzzy-Logic Interface,” The Trans. of the Institute of Electronics and Communication Engineers, Vol.J67-C, No.4, pp. 391-396, 1984 (in English and in Japanese).
  2. K. Uehara and M. Fujise, “Fuzzy Inference Based on Families of α-Level Sets,” IEEE Transactions on Fuzzy Systems, Vol. 1, No. 2, pp. 111–124, 1993.
  3. K. Uehara and M. Fujise, “Multistage Fuzzy Inference Formulated as Linguistic-Truth-Value Propagation and its Learning Algorithm Based on Back-Propagating Error Information,” IEEE Transactions on Fuzzy Systems, Vol. 1, No. 3, pp. 205–221, 1993.
  4. K. Uehara, “Fuzzy Inference Based on a Weighted Average of Fuzzy Sets and its Learning Algorithm for Fuzzy Exemplars,” Proc. of the International Joint Conference of the Fourth IEEE International Conference on Fuzzy Systems and the Second International Fuzzy Engineering Symposium (FUZZ-IEEE/IFES’95), Vol. IV, pp. 2253–2260, 1995.
  5. K. Uehara and K. Hirota, “Fuzzy Connection Admission Control for ATM Networks Based on Possibility Distribution of Cell Loss Ratio,” IEEE Journal on Selected Areas in Communications, Vol. 15, No. 2, pp. 179–190, 1997.
  6. K. Uehara and K. Hirota, “Parallel Fuzzy Inference Based on α-Level Sets and Generalized Means,” Information Sciences (International Journal), Vol. 100, No. 1–4, pp. 165–206, 1997.
  7. K. Uehara and K. Hirota, “Parallel and Multistage Fuzzy Inference Based on Families of α-Level Sets,” Information Sciences (International Journal ), Vol. 106, No. 1-2, pp. 159–195, 1998.
  8. K. Uehara, T. Koyama, and K. Hirota, “Fuzzy Inference with Schemes for Guaranteeing Convexity and Symmetricity in Consequences Based on α-Cuts,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 13, No. 2, pp. 135–149, 2009.
  9. K. Uehara, T. Koyama, and K. Hirota, “Inference with Governing Schemes for Propagation of Fuzzy Convex Constraints Based on α-Cuts,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 13, No. 3, pp. 321–330, 2009.
  10. K. Uehara, T. Koyama, and K. Hirota, “Inference Based on α-Cut and Generalized Mean with Fuzzy Tautological Rules,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 14, No. 1, pp. 76–88, 2010.
  11. K. Uehara, T. Koyama, and K. Hirota, “Suppression Effect of α-Cut Based Inference on Consequence Deviations,&dquo; Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 14, No. 3, pp. 256–271, 2010.
  12. K. Uehara, T. Koyama, and K. Hirota, “Inference Based on α-Cut and Generalized Mean in Representing Fuzzy-Valued Functions,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 14, No. 6, pp. 581–592. 2010.
  13. K. Uehara, S. Sato, and K. Hirota, “Inference for Nonlinear Mapping with Sparse Fuzzy Rules Based on Multi-Level Interpolation,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 15, No. 3, pp. 264–287, 2011.
  14. K. Tsukamoto and K. Uehara, “A Map Matching Method with Fuzzy Inference” IEEJ Transactions on Electronics, Information and Systems, Vol. 132, No. 2, 334-335, 2012 (in Japanese).
  15. K. Uehara and K. Hirota, “Infinite-Level Interpolation for Inference with Sparse Fuzzy Rules: Fundamental Analysis Toward Practical Use,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 17, No. 1, pp. 44–59, 2013.
  16. K. Uehara and K. Hirota, “Multi-Level Interpolation for Inference with Sparse Fuzzy Rules: An Extended Way of Generating Multi-Level Points,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 17, No. 2, pp. 127–148, 2013.
  17. K. Uehara and K. Hirota, “Multi-Level Control of Fuzzy-Constraint Propagation in Inference Based on α-Cuts and Generalized Mean,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 17, No. 4, pp. 647–662, 2013.
  18. K. Uehara and K. Hirota, “Inference with Fuzzy Rule Interpolation at an Infinite Number of Activating Points,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 19, No. 1, pp. 74–90, 2015.
  19. K. Uehara and K. Hirota, “Multi-Level Control of Fuzzy Constraint Propagation via Evaluations with Linguistic Truth Values in Generalized-Mean-Based Inference,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 20, No. 2, pp. 355–377, 2016.
  20. K. Uehara and K. Hirota, “Fuzzy Inference: Its Past and Prospects,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 21, No. 1, pp. 13–19, 2017.
  21. K. Uehara and K. Hirota, “Multi-Level Control of Fuzzy-Constraint Propagation in Inference with Fuzzy Rule Interpolation at an Infinite Number of Activating Points,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 21, No. 3, pp. 425–447, 2017.
  22. K. Uehara and K. Hirota, “Fuzzy Inference Based on α-Cuts and Generalized Mean: Relations Between the Methods in its Family and their Unified Platform,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 21, No. 4, pp. 597–615, 2017.
  23. K. Uehara and K. Hirota, “Noise Reduction with Inference Based on Fuzzy Rule Interpolation at an Infinite Number of Activating Points: Toward fuzzy rule learning in a unified inference platform,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 22, No. 6, pp. 883–899, 2018.
  24. K. Uehara and K. Hirota, “Noise Reduction with Fuzzy Inference Based on Generalized Mean and Singleton Input–Output Rules: Toward fuzzy rule learning in a unified inference platform,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 23, No. 6, pp. 1027–1043, 2019.
  25. K. Uehara and K. Hirota,“Independent Evaluations of Each Fuzzy Rule for Derivative-Free Optimization of Fuzzy Systems: Toward Fast Fuzzy-Rules Learning for Fuzzy Inputs,” Proceedings of the 9th International Symposiumon Computational Intelligence and IndustrialApplications (ISCIIA2020), 1A2-2-4, pp. 1–8, 2020.
  26. K. Uehara and K. Hirota, “A Fast Method for Fuzzy Rules Learning with Derivative-Free Optimization by Formulating Independent Evaluations of Each Fuzzy Rule,” Journal of Advanced Computational Intelligence and Intelligent Informatics, Vol. 25, No. 2, pp. 213–225, 2021.
  27. K. Uehara,“Noise Reduction by Fuzzy Inference Based on α-Cuts and Generalized Mean: A Feasibility Study,” Proceedings of the 7th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2021), M2-6-5, pp. 1–7, 2021.
  28. K. Uehara,“Continual Learning with Fuzzy Inference Based on Derivative-Free Independent Optimization of Each Fuzzy Rule: A Feasibility Study,” Proceedings of Continual Learning and Emergence of Intelligent Systems, Japan Society for Fuzzy Theory and Intelligent Informatics, p. 23, Dec. 2023.
 More papers are listed here.
 

Awards

Award             Journal/Conference Name
(1) JACIII Best Paper Award 2011 Journal of Advanced Computational Intelligence and Intelligent Informatics
(2) ISCIIA2016 Best Paper Award The 7th International Symposiumon Computational Intelligence and Industrial Applications (ISCIIA2016), 2016.
(3) ISCIIA2016 Session Best Presentation Award The 7th International Symposiumon Computational Intelligence and Industrial Applications (ISCIIA2016), 2016.
(4) IWACIII2017 Session Best Presentation Award The 5th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2017), 2017.
(5) ISCIIA-ITCA2018 Session Best Presentation Award The 8th International Symposiumon Computational Intelligence and IndustrialApplications (ISCIIA2018), The 12th China-Japan International Workshop on Information Technology and Control Applications (ITCA2018), 2018.
(6) IWACIII2019 Session Best Presentation Award The 6th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2019), 2019.
(7) JACIII Outstanding Reviewer Award 2019 Journal of Advanced Computational Intelligence and Intelligent Informatics
(8) ISCIIA2020 Session Best Presentation Award The 9th International Symposiumon Computational Intelligence and Industrial Applications (ISCIIA2020), 2020.
 

Main Recent Activities

Year  Official Position       Institute/Conference Name
2021-present Councilor
(評議員)
Japan Society for Fuzzy Theory and Intelligent Informatics
(日本知能情報ファジィ学会)
2019-2021 Director
(理事, 広報委員長)
Japan Society for Fuzzy Theory and Intelligent Informatics
(日本知能情報ファジィ学会)
2004-present Editorial Member Journal of Advanced Computational Intelligence and Intelligent Informatics
2021 Chairperson The 7th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2021)
2021 Program Committee Member The 7th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2021)
2017-2020 Vice Editor-in-Chief for Papers
(論文副委員長)
Editorial Board, Japan Society for Fuzzy Theory and Intelligent Informatics
(日本知能情報ファジィ学会 会誌編集委員会)
2020 Chairperson The 9th International Symposiumon Computational Intelligence and Industrial Applications (ISCIIA2020)
2020 Program Committee Member The 9th International Symposiumon Computational Intelligence and Industrial Applications (ISCIIA2020)
2020 Program Committee Member Joint 11th International Conference on Soft Computing and Intelligent Systems and 21st International Symposium on Advanced Intelligent Systems (SCIS&ISIS2020)
2019 Program Committee Member The 6th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2019)
2018 Program Committee Member The 8th International Symposiumon Computational Intelligence and IndustrialApplications (ISCIIA2018), The 12th China-Japan International Workshop on Information Technology and Control Applications (ITCA2018)
2017 Program Committee Member The 5th International Workshop on Advanced Computational Intelligence and Intelligent Informatics (IWACIII2017)
2016 Program Committee Member The 7th International Symposiumon Computational Intelligence and Industrial Applications (ISCIIA2016)
2017 Associate Editor Journal of Advanced Computational Intelligence and Intelligent Informatics
2013 Guest Editor Journal of Advanced Computational Intelligence and Intelligent Informatics
 More activities are listed here.
 

Profile in Japanese:

上原 清彦
茨城大学大学院  理工学研究科  准教授,     博士(工学)[東京工業大学 授与]

研究分野:ファジィ理論、ファジィ推論、ファジィ論理、ファジィルール学習、ファジィ推論による人工知能、人工免疫システム

職務経験:

Copyright © Kiyohiko Uehara