A constructive algorithm to synthesize arbitrarily connected feedforward neural networks

Puma-Villanueva, Wilfredo J.; Santos, E.P. dos; Zuben, Fernando J. Von

doi:10.1016/j.neucom.2011.05.025

Cited by 25 publications

(10 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Experimental results also con¯rmed that the Algorithm for Optimum Feature Ordering is able to obtain applicable À À À À À À À À À À À À 18.63 3 SVDD 26 À À À À À À À À À À À À 13.0 4 CCA 27 À À À À À À À À À À À À 19.1 5 CMAC NN 28 25.57 3.94 35.22 À À À À À À 6 SGHS 29 À À À 3.00 35.02 6.59 À À À 7 ACFNNA 30 20.87 1.08 À À À 1.58 À À À 8 ISO-FLANN 31 20.37 2.64 À À À À À À À À À 9 RBFN 32 27.30 4.54 30.99 À À À À À À 10 PGFN 32 23.96 6.25 31.16 À À À À À À 11 AT CasPer 33 22.88 1. 29 28.42 1.53 À À À 12 A CasPer 33 23.14 1.15 27.68 1.67 À À À 13 Layered CasPer 33 23.91 2.13 30.38 4.33 À À À 14 pPCA 34 25.00 1.78 32.07 5.87 À À À 15 MDEGL 35 24.84 4.34 37.13 À À À À À À 16 SOCPNN 36 23.71 3.44 52.79 À À À À À À 17 MOCPNN 36 23.71 3.44 37.91 À À À À À À 18 UETS 37 22.38 0.91 À À À 2.46 À À À T. Wang et al…”

Section: Discussionunclassified

Optimized Neural Incremental Attribute Learning for Classification Based on Statistical Discriminability

Wang

Guan

Man

et al. 2014

Int. J. Comp. Intel. Appl.

View full text Add to dashboard Cite

Feature ordering is a signi¯cant data preprocessing method in incremental attribute learning (IAL), where features are gradually trained according to a given order. Previous research showed feature ordering is crucial to the IAL performance. It is relevant to each feature's discrimination ability, which can be calculated by single discriminability (SD). However, when feature dimensions increase, feature discrimination ability should also be calculated incrementally, because discrimination ability in lower dimensional spaces is di®erent from that in higher spaces. Thus based on SD, accumulative discriminability (AD), a new statistical metric for incremental feature discrimination ability estimation, is designed. Moreover, a criterion that summarizes all the produced values of AD is employed to obtain the optimum feature ordering for classi¯cation problems based on neural networks by means of IAL. In addition, in order to reduce the time consumption, an e®ective feature ordering approach is developed. Compared with the feature ordering obtained by other approaches, the method outlined in this paper obtained good¯nal classi¯cation results, which indicates that,¯rstly, feature discrimination ability should be incrementally estimated in IAL; and secondly, feature ordering derived by AD and its corresponding approaches are applicable with IAL.

show abstract

Section: Discussionunclassified

Optimized Neural Incremental Attribute Learning for Classification Based on Statistical Discriminability

Wang

Guan

Man

et al. 2014

Int. J. Comp. Intel. Appl.

View full text Add to dashboard Cite

show abstract

“…In contrast, the simultaneous evolution of architecture and weights is muddled by the well-known permutation problem. Additionally, the computational cost of implementing evolutionary algorithms, even for structure determination alone, may become prohibitive [41].…”

Section: Evolutionary Approachmentioning

confidence: 99%

A New Multilayer Perceptron Pruning Algorithm for Classification and Regression Applications

Thomas

Suhner

2014

Neural Process Lett

View full text Add to dashboard Cite

Optimizing the structure of neural networks remains a hard task. If too small, the architecture does not allow for proper learning from the data, whereas if the structure is too large, learning leads to the well-known overfitting problem. This paper considers this issue, and proposes a new pruning approach to determine the optimal structure. Our algorithm is based on variance sensitivity analysis, and prunes the different types of unit (hidden neurons, inputs, and weights) sequentially. The stop criterion is based on a performance evaluation of the network results from both the learning and validation datasets. Four variants of this algorithm are proposed. These variants use two different estimators of the variance. They are tested and compared with four classical algorithms on three classification and three regression problems. The results show that the proposed algorithms outperform the classical approaches in terms of both computational time and accuracy.

show abstract

“…The authors hypothesize that constructive-pruning hybrid techniques successfully avoid the local optima that hindered previous algorithms. Many such hybrid techniques have been explored [26,105,128,130,209,232,299].…”

Section: Constructive and Pruning Algorithmsmentioning

confidence: 99%

Artificial Neurogenesis: An Introduction and Selective Review

Kowaliw

Bredèche

Chevallier³

et al. 2014

Growing Adaptive Machines

View full text Add to dashboard Cite

International audienceIn this introduction and review—like in the book which follows—we explore the hypothesis that adaptive growth is a means of producing brain-like machines. The emulation of neural development can incorporate desirable characteristics of natural neural systems into engineered designs. The introduction begins with a review of neural development and neural models. Next, artificial development— the use of a developmentally-inspired stage in engineering design—is introduced. Several strategies for performing this " meta-design " for artificial neural systems are reviewed. This work is divided into three main categories: bio-inspired representations ; developmental systems; and epigenetic simulations. Several specific network biases and their benefits to neural network design are identified in these contexts. In particular, several recent studies show a strong synergy, sometimes interchange-ability, between developmental and epigenetic processes—a topic that has remained largely under-explored in the literature

show abstract

A constructive algorithm to synthesize arbitrarily connected feedforward neural networks

Cited by 25 publications

References 21 publications

Optimized Neural Incremental Attribute Learning for Classification Based on Statistical Discriminability

Optimized Neural Incremental Attribute Learning for Classification Based on Statistical Discriminability

A New Multilayer Perceptron Pruning Algorithm for Classification and Regression Applications

Artificial Neurogenesis: An Introduction and Selective Review

Contact Info

Product

Resources

About