Extending the Universal Approximation Theorem for a Broad Class of Hypercomplex-Valued Neural Networks

被引:3
|
作者
Vital, Wington L. [1 ]
Vieira, Guilherme [1 ]
Valle, Marcos Eduardo [1 ]
机构
[1] Univ Estadual Campinas, Campinas, Brazil
来源
INTELLIGENT SYSTEMS, PT II | 2022年 / 13654卷
基金
巴西圣保罗研究基金会;
关键词
Hypercomplex algebras; Neural networks; Universal approximation theorem;
D O I
10.1007/978-3-031-21689-3_45
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The universal approximation theorem asserts that a single hidden layer neural network approximates continuous functions with any desired precision on compact sets. As an existential result, the universal approximation theorem supports the use of neural networks for various applications, including regression and classification tasks. The universal approximation theorem is not limited to real-valued neural networks but also holds for complex, quaternion, tessarines, and Clifford-valued neural networks. This paper extends the universal approximation theorem for a broad class of hypercomplex-valued neural networks. Precisely, we first introduce the concept of non-degenerate hypercomplex algebra. Complex numbers, quaternions, and tessarines are examples of non-degenerate hypercomplex algebras. Then, we state the universal approximation theorem for hypercomplex-valued neural networks defined on a non-degenerate algebra.
引用
收藏
页码:646 / 660
页数:15
相关论文
共 50 条
  • [21] Optimal approximation using complex-valued neural networks
    Geuchen, Paul
    Voigtlaender, Felix
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [22] Quantitative Approximation Results for Complex-Valued Neural Networks
    Caragea, Andrei
    Lee, Dae Gwan
    Maly, Johannes
    Pfander, Goetz
    Voigtlaender, Felix
    [J]. SIAM JOURNAL ON MATHEMATICS OF DATA SCIENCE, 2022, 4 (02): : 553 - 580
  • [23] The Universal Approximation Capabilities of Cylindrical Approximate Identity Neural Networks
    Zarita Zainuddin
    Saeed Panahian Fard
    [J]. Arabian Journal for Science and Engineering, 2016, 41 : 3027 - 3034
  • [24] Universal approximation of flows of control systems by recurrent neural networks
    Aguiar, Miguel
    Das, Amritam
    Johansson, Karl H.
    [J]. 2023 62ND IEEE CONFERENCE ON DECISION AND CONTROL, CDC, 2023, : 2320 - 2327
  • [25] Conditions for RBF Neural Networks to Universal Approximation and Numerical Experiments
    Nong, Jifu
    [J]. COMMUNICATIONS AND INFORMATION PROCESSING, PT 2, 2012, 289 : 299 - 308
  • [26] On the universal approximation property of radial basis function neural networks
    Ismayilova, Aysu
    Ismayilov, Muhammad
    [J]. ANNALS OF MATHEMATICS AND ARTIFICIAL INTELLIGENCE, 2024, 92 (03) : 691 - 701
  • [27] The Universal Approximation Capabilities of Cylindrical Approximate Identity Neural Networks
    Zainuddin, Zarita
    Fard, Saeed Panahian
    [J]. ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING, 2016, 41 (08) : 3027 - 3034
  • [28] Performance of Deep and Shallow Neural Networks, the Universal Approximation Theorem Activity Cliffs, and QSAR (vol 36, 1600118, 2017)
    Winkler, David A.
    Le, Tu C.
    [J]. MOLECULAR INFORMATICS, 2017, 36 (11)
  • [29] Approximation of Fuzzy Neural Networks to Fuzzy-Valued Measurable Function
    Li, Jun
    Gao, Shuai
    Chen, Bin
    [J]. QUANTITATIVE LOGIC AND SOFT COMPUTING 2010, VOL 2, 2010, 82 : 351 - 358
  • [30] Universal Approximation Property and Equivalence of Stochastic Computing-Based Neural Networks and Binary Neural Networks
    Wang, Yanzhi
    Zhan, Zheng
    Zhao, Liang
    Tang, Jian
    Wang, Siyue
    Li, Jiayu
    Yuan, Bo
    Wen, Wujie
    Lin, Xue
    [J]. THIRTY-THIRD AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FIRST INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE / NINTH AAAI SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2019, : 5369 - 5376