Abstract
Online sequential extreme learning machine (OS-ELM) is one of the most popular real-time learning strategy for feedforward neural networks with single hidden layer due to its fast learning speed and excellent generalization ability. When dealing with quaternion signals, traditional real-valued learning models usually provide only suboptimal solutions compared with their quaternion-valued counterparts. However, online sequential quaternion extreme learning machine (OS-QELM) model is still lacking in literature. To fill this gap, this paper aims to establish a framework for the derivation and the design of OS-QELM. Specifically, we first derive a standard OS-QELM, and then propose two augmented OS-QELM models which can capture the complete second-order statistics of noncircular quaternion signals. The corresponding regularized models and two approaches to reducing the computational complexity are also derived and discussed respectively. Benefiting from the quaternion algebra and the augmented structure, the proposed models exhibit superiority over OS-ELM in simulation results on several benchmark quaternion regression problems and colour face recognition problems.
Similar content being viewed by others
References
Li MB, Huang GB, Zhu QY, Siew CK (2006) Extreme learning machine: theory and applications. Neurocomputing 70:489–501
Huang GB, Zhou H, Ding X, Zhang R (2012) Extreme learning machine for regression and multiclass classification. IEEE Trans Syst Man Cybern B Cybern 42(2):513-529
Grasso F, Luchetta A, Manetti S (2018) A multi-valued neuron based complex ELM neural network. Neural Process Lett 48(1):389–401
Ren Z, Yang L (2019) Robust extreme learning machines with different loss functions. Neural Process Lett 49(3):1543–1565
Chang P, Zhang J, Hu J, Song Z (2018) A deep neural network based on ELM for semi-supervised learning of image classification. Neural Process Lett 48(1):375–388
Strogatz SH (2001) Nonlinear dynamics and chaos: with applications to physics, biology, chemistry and engineering (studies in nonlinearity). Westview Press, Boulder
Wang J, Zhang BJ, Sang ZY, Liu YS, Wu SJ, Miao Q (2020) Convergence of a modified gradient-based learning algorithm with penalty for single-hidden-layer feed-forward networks. Neural Comput Appl 32:2445–2456
Zhang HS, Wu W, Liu F, Yao MC (2009) Boundedness and convergence of online gradient method with penalty for feedforward neural networks. IEEE Trans Neural Netw 20(6):1050–1054
Sun M, Bai Q (2011) A new descent memory gradient method and its global convergence. J Syst Sci Complex 24(4):784
Zhang HS, Mandic DP (2016) Is a complex-valued stepsize advantageous in complex-valued gradient learning algorithms? IEEE Trans Neural Netw Learn Syst 27(12):2730–2735
Li F, Zurada JM, Liu Y, Wu W (2017) Input layer regularization of multilayer feedforward neural networks. IEEE Access 5:10979–10985
Huang GB, Zhu QY, Siew CK (2015) Trends in extreme learning machines: a review. Neural Netw 61:32–48
Yang DK, Li ZX, Wu W (2014) Extreme learning machine for interval neural networks. Neural Comput Appl 27:3–8
Qu YP, Shang CJ, Wu W, Shen Q (2011) Evolutionary fuzzy extreme leanrning machine for mammographic risk analysis. Int J Fuzzy Syst 13(4):282–291
Fan YT, Wu W, Yang WY, Fan QW, Wang J (2014) A Pruning algorithm with \(l_{1/2}\) regularizer for extreme learning machine. J Zhejiang Univ (Eng Sci) 15:119–25
Yang S, Zhang C, Bao Y, Yang J, Wu W (2020) Binary output layer of extreme learning machine for solving multi-class classification problems. Neural Process Lett 52(1):1–15
Yang Y, Wang Y, Yuan X (2013) Parallel chaos search based incremental extreme learning machine. Neural Process Lett 37(3):277–301
Liang NY, Huang GB, Saratchandran P, Sundararajan N (2006) A fast and accurate online sequential learning algorithm for feedforward networks. IEEE Trans Neural Netw 17(6):1411–1423
Goldman R (2011) Understanding quaternions. Graph Models 73(2):21–49
Morais JP, Georgiev S, Sprößig W (2014) Real quaternionic calculus handbook. Springer, Basel
Tobar FA, Mandic DP (2014) Quaternion reproducing kernel Hilbert spaces: existence and uniqueness conditions. IEEE Trans Inf Theory 60(9):5736–5749
Xia YL, Jahanchahi C, Nitta T, Mandic DP (2015) Performance bounds of quaternion estimators. IEEE Trans Neural Netw Learn Syst 26(12):3287–3292
Lian D, Xu L, Jiang H (2005) The properties of similar quaternions and their analytic application. J Nat Sci Nanjing Norm Univ (in Chinese) 2:10–15
Mitsubori K, Saito T (1994) Torus doubling and hyperchaos in a five dimensional hysteresis circuit. Proc IEEE Int Symp Circ Syst 6:113–116
Liu Z, Qiu Y, Peng Y, Pu J, Zhang X (2017) Quaternion based maximum margin criterion method for color face recognition. Neural Process Lett 45(3):913–923
Arena P, Fortuna L, Muscato G, Xibilia MG (1998) Neural networks in multidimensional domains. Lecture Notes in Control and Information Sciences 234
Arena P, Fortuna L, Muscatoandm G, Xibilia G (1997) Multilayer perceptrons to approximate quaternion valued functions. Neural Netw 10(2):335–342
Greenblatt AB, Agaian SS (2018) Introducing quaternion multi-valued neural networks with numerical examples. Inf Sci 423:326–342
Mandic DP, Goh SL (2009) Complex valued nonlinear adaptive filters: noncircularity, widely linear and neural models. Wiley, New York
Took CC, Mandic DP (2009) The quaternion LMS algorithm for adaptive filtering of hypercomplex processes. IEEE Trans Signal Process 57(4):1316–1327
Ujang BC, Took CC, Mandic DP (2011) Quaternion-valued nonlinear adaptive filtering. IEEE Trans Neural Netw 22:1193–1206
Jahanchahi C, Mandic DP (2014) A class of quaternion Kalman filters. IEEE Trans Neural Netw Learn Syst 25(3):533–544
Xiang M, Kanna S, Mandic DP (2018) Performance analysis of quaternion-valued adaptive filters in nonstationary environments. IEEE Trans Signal Process 66(6):1566–1579
Xu DP, Mandic DP (2015) The theory of quaternion matrix derivatives. IEEE Trans Signal Process 63(6):1543–1556
Xu DP, Jahanchahi C, Cheong Took C, Mandic DP (2015) Enabling quaternion derivatives: the generalized HR calculus. R Soc Open Sci 2(8):150–255
Xia YL, Jahanchahi C, Mandic DP (2015) Quaternion-valued echo state networks. IEEE Trans Neural Netw Learn Syst 26(4):663–673
Shi ZJ, Guo J (2009) A new family of conjugate gradient methods. J Comput Appl Math 224(1):444–457
Mandic DP, Jahanchahi C, Took CC (2011) A quaternion gradient operator and its applications. IEEE Signal Proc Lett 18:47–50
Javidi S, Took CC, Mandic DP (2011) Fast independent component analysis algorithm for quaternion valued signals. IEEE Trans. Neural Netw 22(12):1967–1978
Zhang HS, Lv H (2019) Augmented quaternion extreme learning machine. IEEE Access 7:90842–90850
Cheong Took C, Mandic DP (2011) Augmented second order statistics of quaternion random signals. Signal Process 91(2):214–224
Xia YL, Douglas SC, Mandic DP (2018) A perspective on clms as a deficient length augmented clms: dealing with second order noncircularity. Signal Process 149:236–245
Xia YL, Mandic DP (2017) Complementary mean square analysis of augmented CLMS for second-order noncircular Gaussian signals. IEEE Signal Process Lett 24(9):1413–1417
Took CC, Mandic DP (2010) A quaternion widely linear adaptive filter. IEEE Trans Signal Process 58(8):4427–4431
Picinbono B, Chevalier P (1995) Widely linear estimation with complex data. IEEE Trans Signal Process 43(8):2030–2033
Zhang HS, Wang YY, Xu DP, Wang J, Xu LH (2018) The augmented complex-valued extreme learning machine. Neurocomputing 311:363–372
Via J, Ramirez D, Santamaria I (2010) Properness and widely linear processing of quaternion random vectors. IEEE Trans Inf Theory 56(7):3502–3515
Xia YL, Mandic DP (2017) Augmented performance bounds on strictly linear and widely linear estimators with complex data. IEEE Trans Signal Process 66(2):507–514
Xia YL, Tao S, Li Z, Xiang M, Pei W, Mandic DP (2019) Full mean square performance bounds on quaternion estimators for improper data. IEEE Trans Signal Process 67(15):4093–4106
Ward JP (2012) Quaternions and Cayley numbers: algebra and applications, vol 403. Springer, Berlin
Ell TA, Sangwine SJ (2007) Quaternion involutions and anti-involutions. Comput Math Appl 53(1):137–143
Liu B (2004) The Moore-Penrose generalized inverse of quaternion matrix. J Guilin Univ Electr Tech 24(5):68–71
Golub GH, Loan CFV (1996) Matrix computations, 3rd edn. The Johns Hopkins Univ. Press, Baltimore
Bartlett L (1998) The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network. IEEE Trans. Inform. Theory 44(2):525–536
Chong EKP, Zak SH (2001) An introduction to optimization. Wiley, New York
Kirkpatrick J, Pascanu R, Rabinowitz N et al (2017) Overcoming catastrophic forgetting in neural networks. Proc Natl Acad Sci 114(13):3521–3526
Jahanchahi C, Took CC, Mandic DP (2010) The widely linear quaternion recursive least squares filter. In: 2010 2nd international workshop on cognitive information processing. IEEE. 87-92
Xiang M, Took CC, Mandic DP (2017) Cost-effective quaternion minimum mean square error estimation: from widely linear to four-channel processing. Signal Process 136:81–91
Spacek L (2009) Libor Spacek’s facial images databases [Online]. available: http://cmp.felk.cvut.cz/~spacelib/faces
Jaha ES, Ghouti L (2011) Color face recognition using quaternion PCA. In: Proc. 4th Int. Conf. Imag. Crim. Detect. Prev. 1-6
Acknowledgements
This work is supported by the National Natural Science Foundation of China (Nos. 61671099, 61301202) and the Fundamental Research Funds for the Central Universities of China.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Zhu, S., Wang, H., Lv, H. et al. Augmented Online Sequential Quaternion Extreme Learning Machine. Neural Process Lett 53, 1161–1186 (2021). https://doi.org/10.1007/s11063-021-10435-8
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11063-021-10435-8