[1]陈德旺,吴奕然,欧纪祥,等.面向下一代人工智能的三代模糊系统的发展与展望[J].智能系统学报,2026,21(3):566-576.[doi:10.11992/tis.202506002]
CHEN Dewang,WU Yiran,OU Jixiang,et al.Development and prospects of third-generation fuzzy systems for next-generation artificial intelligence[J].CAAI Transactions on Intelligent Systems,2026,21(3):566-576.[doi:10.11992/tis.202506002]
点击复制
CHEN Dewang,WU Yiran,OU Jixiang,et al.Development and prospects of third-generation fuzzy systems for next-generation artificial intelligence[J].CAAI Transactions on Intelligent Systems,2026,21(3):566-576.[doi:10.11992/tis.202506002]
面向下一代人工智能的三代模糊系统的发展与展望
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
21
期数:
2026年第3期
页码:
566-576
栏目:
综述
出版日期:
2026-05-05
- Title:
- Development and prospects of third-generation fuzzy systems for next-generation artificial intelligence
- Keywords:
- next-generation artificial intelligence; fuzzy systems; deep learning; optimization; big data; neural networks; rule-based systems; interpretablity
- 分类号:
- TP39
- 文献标志码:
- 2026-3-13
- 摘要:
- 可解释、可通用的下一代人工智能方法的重大研究计划旨在回应国家在人工智能发展中的战略需求，聚焦人工智能的基础科学问题，推动新型方法体系的构建。与内部机制难以解释的神经网络算法相比，模糊系统因其较强的可解释性与良好的鲁棒性，在不确定性问题建模与决策中发挥重要作用。模糊系统的发展历程可分为3个阶段：基于专家经验的传统模糊系统、面向低维小数据的自适应模糊系统、面向高维大数据的深度优化模糊系统。当前的研究重点在于第三代模糊系统，即深度优化模糊系统。该系统通过分层结构设计、规则自适应生成机制以及与进化算法、梯度下降等优化技术的深度融合，能够从海量高维数据中自动提取有效特征与潜在规则，并在逐层抽象过程中实现特征的降维与重构。由此，深度优化模糊系统不仅有效缓解了传统模糊系统在高维环境下面临的“规则爆炸”问题，显著提升了计算效率与泛化能力，实现了可解释性与高精度的有机平衡，也有望为人工智能的未来发展开辟新的方向。
- Abstract:
- The Grand Research Initiative for Explainable and Universally Applicable Next-Generation Artificial Intelligence (AI) Methods, aligned with national strategic priorities for AI development, focuses on fundamental scientific challenges in AI. This initiative aims to develop a new generation of AI methodologies, addressing the lack of interpretability often associated with neural networks. Compared with neural network algorithms whose internal mechanisms are difficult to explain, fuzzy systems offer strong interpretability and robustness and play an important role in modeling and decision-making under uncertainty. The development of fuzzy systems has progressed through three stages: traditional fuzzy systems based on expert experience, adaptive fuzzy systems for low-dimensional small data, and deep optimization fuzzy systems for high-dimensional big data. Current research focuses on the third-generation fuzzy systems, that is, deep optimization fuzzy systems. Through hierarchical architecture, adaptive rule generation mechanisms, and deep integration with optimization techniques such as evolutionary algorithms and gradient descent, these systems can automatically extract effective features and latent rules from massive high-dimensional data and achieve feature dimensionality reduction and reconstruction through layer-by-layer abstraction. Consequently, deep optimization fuzzy systems effectively mitigate the “rule explosion” problem encountered by traditional fuzzy systems in high-dimensional settings, significantly improve computational efficiency and generalization capability, and achieve a balance between interpretability and high accuracy. This is expected to open up a new direction for the future development of artificial intelligence.
- 参考文献/References:
-
[1] 国务院. 国务院关于印发新一代人工智能发展规划的通知: 国发〔2017〕35号[EB/OL]. (2017-07-20)[2025-01-01]. https://www.gov.cn/zhengce/content/2017-07/20/content_5211996.htm. State Council. Notice of the State Council on Issuing the Development Plan on the New Generation of Artificial Intelligence: Guo Fa [2017] No. 35[EB/OL]. (2017-07-20)[2025-01-01]. https://www.gov.cn/zhengce/content/2017-07/20/content_5211996.htm.
[2] 国家自然科学基金委员会. 可解释、可通用的下一代人工智能方法重大研究计划2023年度项目指南[J]. 模式识别与人工智能, 2023, 36(3): 280-284 National Natural Science Foundation of China. Explanatory and universal major research program of the next generation artificial intelligence method 2023 project guide[J]. Pattern recognition and artificial intelligence, 2023, 36(3): 280-284
[3] 清华大学探臻科技评论社, 编. 下一代创新科技: 第1辑[M]. 北京: 清华大学出版社, 2024: 41-52. Tsinghua University TANZEN Science and Technology Review, ed. Next Generation Innovation Technology: Volume 1[M]. Beijing: Tsinghua University Press, 2024: 41-52.
[4] ZADOR A, ESCOLA S, RICHARDS B, et al. Catalyzing next-generation artificial intelligence through NeuroAI[J]. Nature communications, 2023, 14: 1597
[5] 晏勇, 杜继宏, 冯元琨. 模糊控制[J]. 计算机测量与控制, 1999, 7(1): 55-57 YAN Yong, DU Jihong, FENG Yuankun. Fuzzy control[J]. Computer measurement & control, 1999, 7(1): 55-57
[6] BELLMAN R. Decision-Makking in a fuzzy environment[J]. Management science, 1970, 17: 8141-8164
[7] ZADEH L A. Fuzzy sets[J]. Information and control, 1965, 8(3): 338-353
[8] LU Jie, MA Guangzhi, ZHANG Guangquan. Fuzzy machine learning: a comprehensive framework and systematic review[J]. IEEE transactions on fuzzy systems, 2024, 32(7): 3861-3878
[9] ZADEH L A. Probability measures of fuzzy events[J]. Journal of mathematical analysis and applications, 1968, 23(2): 421-427
[10] ZADEH L A. Similarity relations and fuzzy orderings[J]. Information sciences, 1971, 3(2): 177-200
[11] ZADEH L A. Outline of a new approach to the analysis of complex systems and decision processes[J]. IEEE transactions on systems, man, and cybernetics, 1973, SMC-3(1): 28-44.
[12] ISHIBUCHI H, FUJIOKA R, TANAKA H. Neural networks that learn from fuzzy if-then rules[J]. IEEE transactions on fuzzy systems, 1993, 1(2): 85-97
[13] MAMDANI E H, ASSILIAN S. An experiment in linguistic synthesis with a fuzzy logic controller[J]. International journal of man-machine studies, 1975, 7(1): 1-13
[14] YASUNOBU S, MIYAMOTO S, IHARA H. Fuzzy control for automatic train operation system[J]. IFAC proceedings volumes, 1983, 16(4): 33-39
[15] HIROTA K, SUGENO M. Industrial Applications of Fuzzy Technology in the World[M]. Elsevier Science Inc, 1985.
[16] WANG L X. A Course in fuzzy systems and control[M]. Upper Saddle River: Prentice Hall PTR, 1997: 1-424.
[17] WANG L X. Fuzzy systems are universal approximators[C]// IEEE International Conference on Fuzzy Systems. Piscataway: IEEE, 2002: 1163-1170.
[18] WANG L X, MENDEL J M. Fuzzy basis functions, universal approximation, and orthogonal least-squares learning[J]. IEEE transactions on neural networks, 1992, 3(5): 807-814
[19] WANG L X, MENDEL J M. Generating fuzzy rules by learning from examples[J]. IEEE transactions on systems, man, and cybernetics, 1992, 22(6): 1414-1427
[20] ZHAI Yanwei, LYU Zheng, ZHAO Jun, et al. Data-driven inference modeling based on an on-line Wang-Mendel fuzzy approach[J]. Information sciences, 2021, 551: 113-127
[21] 项长生, 王旭, 赵华, 等. 基于融合WM算法的Mamdani型模糊推理的板桥损伤定量分析[J]. 中国安全生产科学技术, 2024, 20(8): 173-180 XIANG Changsheng, WANG Xu, ZHAO Hua, et al. Quantitative analysis of slab bridge damage based on Mamdani fuzzy reasoning with fusion of WM algorithm[J]. Journal of safety science and technology, 2024, 20(8): 173-180
[22] WANG Yuangang, LIU Haoran, JIA Wenjuan, et al. Deep fuzzy rule-based classification system with improved Wang-Mendel method[J]. IEEE transactions on fuzzy systems, 2022, 30(8): 2957-2970
[23] JANG J S R. ANFIS: adaptive-network-based fuzzy inference system[J]. IEEE transactions on systems, man, and cybernetics, 1993, 23(3): 665-685
[24] KOSKO B. Neural networks and fuzzy systems[M]. Englewood Ciffs: Prentice Hall, 1992.
[25] KARNIK N N, MENDEL J M, LIANG Qilian. Type-2 fuzzy logic systems[J]. IEEE transactions on fuzzy systems, 1999, 7(6): 643-658
[26] 陈守煜, 李庆国. 一种新的模糊聚类神经网络及其在水资源评价中的应用[J]. 水利学报, 2005, 36(6): 662-666 CHEN Shouyu, LI Qingguo. Fuzzy clustering neural network and its application to water resources assessment[J]. Journal of hydraulic engineering, 2005, 36(6): 662-666
[27] 袁小芳, 王耀南, 孙炜. 支持向量机-模糊推理自学习控制器设计[J]. 控制理论与应用, 2006, 23(1): 1-6 YUAN Xiaofang, WANG Yaonan, SUN Wei. Self-learning controller using support vector machines and fuzzy inference system[J]. Control theory & applications, 2006, 23(1): 1-6
[28] 马铭. 基于数据驱动的模糊系统建模方法研究[D]. 长春: 吉林大学, 2006. MA Ming. Research on data-driven fuzzy system modeling[D]. Changchun: Jilin University, 2006.
[29] 陈德旺, 蔡际杰, 黄允浒. 面向可解释性人工智能与大数据的模糊系统发展展望[J]. 智能科学与技术学报, 2019, 1(4): 327-334 CHEN Dewang, CAI Jijie, HUANG Yunhu. Development prospect of fuzzy system oriented to interpretable artificial intelligence and big data[J]. Chinese journal of intelligent science and technology, 2019, 1(4): 327-334
[30] 赵文迪, 陈德旺, 卓永强, 等. 深度神经模糊系统算法及其回归应用[J]. 自动化学报, 2020, 46(11): 2350-2358 ZHAO Wendi, CHEN Dewang, ZHUO Yongqiang, et al. Deep neural fuzzy system algorithm and its regression application[J]. Acta automatica sinica, 2020, 46(11): 2350-2358
[31] HUANG Yunhu, CHEN Dewang, ZHAO Wendi, et al. Fuzzy C-means clustering based deep patch learning with improved interpretability for classification problems[J]. IEEE access, 2022, 10: 49873-49891
[32] HUANG Yunhu, LIN Geng, CHEN Dewang, et al. Deep neural-fuzzy system algorithms with improved interpretability for classification problems[J]. International journal of fuzzy systems, 2024, 26(3): 900-921
[33] ZHAO Wendi, CHEN Dewang, ZHENG Xiaoyu, et al. Serial fuzzy system algorithm for predicting biological activity of anti-breast cancer compounds[J]. Applied intelligence, 2023, 53(11): 13801-13814
[34] ZHOU Ta, CHUNG F L, WANG Shitong. Deep TSK fuzzy classifier with stacked generalization and triplely concise interpretability guarantee for large data[J]. IEEE transactions on fuzzy systems, 2017, 25(5): 1207-1221
[35] 王士同, 谢润山, 周尔昊. 可解释的深度TSK模糊系统综述[J]. 数据采集与处理, 2022, 37(5): 935-951 WANG Shitong, XIE Runshan, ZHOU Erhao. Survey of interpretable deep TSK fuzzy systems[J]. Journal of data acquisition and processing, 2022, 37(5): 935-951
[36] GHADIRI N, GHADIRI A, SHEIKHOLESLAMI A. A fuzzy deep learning approach to health-related text classification[C]//Intelligent and Fuzzy Techniques for Emerging Conditions and Digital Transformation. Cham: Springer, 2022: 179-186.
[37] 卞则康, 张进, 王士同. 面向不平衡数据的深度TSK模糊分类器[J]. 模式识别与人工智能, 2023, 36(3): 211-224 BIAN Zekang, ZHANG Jin, WANG Shitong. A deep Takagi-Sugeno-Kang fuzzy classifier for imbalanced data[J]. Pattern recognition and artificial intelligence, 2023, 36(3): 211-224
[38] TALPUR N, ABDULKADIR S J, ALHUSSIAN H, et al. Deep neuro-fuzzy system application trends, challenges, and future perspectives: a systematic survey[J]. Artificial intelligence review, 2023, 56(2): 865-913
[39] 王园园, 史东辉, 甘书灵. 基于深度神经模糊系统的交通事故严重程度预测研究[J]. 软件工程, 2024, 27(8): 62-65,78 WANG Yuanyuan, SHI Donghui, GAN Shuling. Research on predicting the traffic accident severity based on deep neural fuzzy system[J]. Software engineer, 2024, 27(8): 62-65, 78
[40] 施奇环, 张雄涛. 融合深浅层次知识的自学习TSK模糊癫痫辅助检测算法[J]. 模式识别与人工智能, 2025, 38(1): 82-93 SHI Qihuan, ZHANG Xiongtao. Self-learning TSK fuzzy epilepsy assistant detection algorithm incorporating shallow and deep knowledge[J]. Pattern recognition and artificial intelligence, 2025, 38(1): 82-93
[41] FENG Qiying, CHEN Long, CHEN C L P, et al. Deep fuzzy clustering: a representation learning approach[J]. IEEE transactions on fuzzy systems, 2020, 28(7): 1420-1433
[42] WU Dongrui, KING J T, CHUANG C H, et al. Spatial filtering for EEG-based regression problems in brain-computer interface (BCI)[J]. IEEE transactions on fuzzy systems, 2018, 26(2): 771-781
[43] CHEN Cheng, CAO Yu, CHEN Xinxing, et al. A fused deep fuzzy neural network controller and its application to pneumatic flexible joint[J]. IEEE/ASME transactions on mechatronics, 2023, 28(6): 3214-3225
[44] TALPUR N, ABDULKADIR S J, ALHUSSIAN H, et al. A comprehensive review of deep neuro-fuzzy system architectures and their optimization methods[J]. Neural computing and applications, 2022, 34(3): 1837-1875
[45] LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553): 436-444
[46] UTTERBACK J M. Mastering the dynamics of innovation: how companies can seize opportunities in the face of technological change[M]. Boston: Harvard Business School Press, 1994: 253.
[47] FENN J, RASKINO M. Mastering the hype cycle[M]. Boston, Mass.: Harvard Business Press, 2008.
[48] 陈德旺, 王蕊, 孔令坤, 等. 基于模糊系统的第三代人工智能[J]. 智能系统学报, 2025, 20(5): 1071-1081 CHEN Dewang, WANG Rui, KONG Lingkun, et al. Third-generation artificial intelligence based on fuzzy systems[J]. CAAI transactions on intelligent systems, 2025, 20(5): 1071-1081
[49] 陈德旺, 刘俐俐, 赵文迪, 等. 基于模糊系统的定性与定量知识的综合集成[J]. 智能科学与技术学报, 2024, 6(4): 445-455 CHEN Dewang, LIU Lili, ZHAO Wendi, et al. Qualitative and quantitative knowledge of metasynthesis based on fuzzy system[J]. Chinese journal of intelligent science and technology, 2024, 6(4): 445-455
- 相似文献/References:
-
[1]孙富春,杨 晋,刘华平.SISO Mamdani模糊系统作为函数逼近器的必要条件[J].智能系统学报,2009,4(4):288.
SUN Fu-chun,YANG Jin,LIU Hua-ping.Preconditions for SISO Mamdani fuzzy systems to perform as function approximators[J].CAAI Transactions on Intelligent Systems,2009,4():288.
[2]蒋云良,印泽宗,张雄涛,等.高阶Takagi-Sugeno-Kang模糊知识蒸馏分类器及其在脑电信号分类中的应用[J].智能系统学报,2024,19(6):1419.[doi:10.11992/tis.202307029]
JIANG Yunliang,YIN Zezong,ZHANG Xiongtao,et al.TSK fuzzy distillation classifier with negative Euclidean probability and High-order fuzzy dark knowledge transfer and its application on EEG signals classification[J].CAAI Transactions on Intelligent Systems,2024,19():1419.[doi:10.11992/tis.202307029]
[3]陈德旺,王蕊,孔令坤,等.基于模糊系统的第三代人工智能[J].智能系统学报,2025,20(5):1071.[doi:10.11992/tis.202407011]
CHEN Dewang,WANG Rui,KONG Lingkun,et al.Third-generation artificial intelligence based on fuzzy systems[J].CAAI Transactions on Intelligent Systems,2025,20():1071.[doi:10.11992/tis.202407011]
[4]张雄涛,陈天宇,赵康,等.基于多教师自适应知识蒸馏的TSK模糊分类器[J].智能系统学报,2025,20(5):1136.[doi:10.11992/tis.202410028]
ZHANG Xiongtao,CHEN Tianyu,ZHAO Kang,et al.TSK fuzzy classifier based on multi-teacher adaptive knowledge distillation[J].CAAI Transactions on Intelligent Systems,2025,20():1136.[doi:10.11992/tis.202410028]
备注/Memo
收稿日期:2025-6-3。
基金项目:国家自然科学基金项目(62461160259)；福建省闽江学者讲座教授人才计划项目(GY-Z24014)；福建第三批创新之星人才计划(003002)；福建省财政厅教育科研专项(GY-Z21001)；福建理工大学科研基金项目(GY-Z22071).
作者简介:陈德旺，教授，博士生导师，IET Fellow，福建省“闽江学者”、福建省创新之星，中国自动化学会计算智能及其应用专委会主任，主要研究方向为人工智能算法、模糊系统、智能交通系统。发表学术论文200余篇，总被引超4000 余次。E-mail：dwchen@fjut.edu.cn。;吴奕然，硕士研究生，主要研究方向为可解释人工智能。E-mail：macunwy@163.com。;熊刚，研究员，博士，主要研究方向为复杂系统平行控制、智能交通、智能制造。E-mail：gang.xiong@ia.ac.cn。
通讯作者:熊刚. E-mail：gang.xiong@ia.ac.cn
更新日期/Last Update:
1900-01-01