Auto-association by multilayer perceptrons and singular value decomposition - PubMed (original) (raw)
Auto-association by multilayer perceptrons and singular value decomposition
H Bourlard et al. Biol Cybern. 1988.
Abstract
The multilayer perceptron, when working in auto-association mode, is sometimes considered as an interesting candidate to perform data compression or dimensionality reduction of the feature space in information processing applications. The present paper shows that, for auto-association, the nonlinearities of the hidden units are useless and that the optimal parameter values can be derived directly by purely linear techniques relying on singular value decomposition and low rank matrix approximation, similar in spirit to the well-known Karhunen-Loève transform. This approach appears thus as an efficient alternative to the general error back-propagation algorithm commonly used for training multilayer perceptrons. Moreover, it also gives a clear interpretation of the rôle of the different parameters.
Similar articles
- A learning rule for very simple universal approximators consisting of a single layer of perceptrons.
Auer P, Burgsteiner H, Maass W. Auer P, et al. Neural Netw. 2008 Jun;21(5):786-95. doi: 10.1016/j.neunet.2007.12.036. Epub 2007 Dec 31. Neural Netw. 2008. PMID: 18249524 - Determining and improving the fault tolerance of multilayer perceptrons in a pattern-recognition application.
Emmerson MD, Damper RI. Emmerson MD, et al. IEEE Trans Neural Netw. 1993;4(5):788-93. doi: 10.1109/72.248456. IEEE Trans Neural Netw. 1993. PMID: 18276508 - [Networks of excitable elements. VIII. Methods of approximation of the theory of nets].
Chetaev AN. Chetaev AN. Biofizika. 1976 Sep-Oct;21(5):910-2. Biofizika. 1976. PMID: 1022253 Russian. No abstract available. - Understanding the mind of a worm: hierarchical network structure underlying nervous system function in C. elegans.
Chatterjee N, Sinha S. Chatterjee N, et al. Prog Brain Res. 2008;168:145-53. doi: 10.1016/S0079-6123(07)68012-1. Prog Brain Res. 2008. PMID: 18166392 Review. - Emerging principles governing the operation of neural networks.
Getting PA. Getting PA. Annu Rev Neurosci. 1989;12:185-204. doi: 10.1146/annurev.ne.12.030189.001153. Annu Rev Neurosci. 1989. PMID: 2648949 Review. No abstract available.
Cited by
- Comparative Performance of Autoencoders and Traditional Machine Learning Algorithms in Clinical Data Analysis for Predicting Post-Staged GKRS Tumor Dynamics.
Volovăț SR, Popa TO, Rusu D, Ochiuz L, Vasincu D, Agop M, Buzea CG, Volovăț CC. Volovăț SR, et al. Diagnostics (Basel). 2024 Sep 21;14(18):2091. doi: 10.3390/diagnostics14182091. Diagnostics (Basel). 2024. PMID: 39335770 Free PMC article. - Compressed representation of brain genetic transcription.
Ruffle JK, Watkins H, Gray RJ, Hyare H, Thiebaut de Schotten M, Nachev P. Ruffle JK, et al. Hum Brain Mapp. 2024 Aug 1;45(11):e26795. doi: 10.1002/hbm.26795. Hum Brain Mapp. 2024. PMID: 39045881 Free PMC article. - Adaptive habitat biogeography-based optimizer for optimizing deep CNN hyperparameters in image classification.
Xin J, Khishe M, Zeebaree DQ, Abualigah L, Ghazal TM. Xin J, et al. Heliyon. 2024 Mar 21;10(7):e28147. doi: 10.1016/j.heliyon.2024.e28147. eCollection 2024 Apr 15. Heliyon. 2024. PMID: 38689992 Free PMC article. - Artificial Neural Networks for Short-Form Development of Psychometric Tests: A Study on Synthetic Populations Using Autoencoders.
Casella M, Dolce P, Ponticorvo M, Milano N, Marocco D. Casella M, et al. Educ Psychol Meas. 2024 Feb;84(1):62-90. doi: 10.1177/00131644231164363. Epub 2023 Apr 15. Educ Psychol Meas. 2024. PMID: 38250505 - Semi-2DCAE: a semi-supervision 2D-CNN AutoEncoder model for feature representation and classification of encrypted traffic.
Cui J, Bai L, Li G, Lin Z, Zeng P. Cui J, et al. PeerJ Comput Sci. 2023 Nov 9;9:e1635. doi: 10.7717/peerj-cs.1635. eCollection 2023. PeerJ Comput Sci. 2023. PMID: 38077604 Free PMC article.
References
- J Acoust Soc Am. 1988 Apr;83(4):1615-26 - PubMed
MeSH terms
LinkOut - more resources
Other Literature Sources