Scaled Conjugate Gradient Learning for Quaternion-Valued Neural Networks

Popa, Călin-Adrian

doi:10.1007/978-3-319-46675-0_27

Cited by 6 publications

(5 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is an extension to enhance performance of back propagation algorithm. This method applied on real world applications show improved results over complex gradient descent algorithm (Popa, 2015).…”

Section: Automated Detection and Classification Of Microcalcificationmentioning

confidence: 92%

Automated Detection and Classification of Microcalcification Clusters with Enhanced Preprocessing and Fractal Analysis

Gowri

Valluvan²,

Chamundeeswari³

2018

Asian Pac J Cancer Prev

View full text Add to dashboard Cite

This paper addresses the automated detection of microcalcification clusters from mammogram images by enhanced preprocessing operations on digital mammograms for automated extraction of breast tissue from background, removing artefacts occurring during image registration using X-rays, followed by fractal analysis of suspicious regions. Identification of breast of either left or right and realigning them to a standard position forms a primitive step in preprocessing of mammograms. As the next step in the process, pectoral muscles are separated. Suspicious regions of microcalcifications are identified and are subjected to further analysis of classifying it as benign or malignant. Texture features are representative of its malignancy and fractal analysis was carried out on extracted suspicious regions for its texture features. Principal Component Analysis was carried out to extract optimal features. Ten features were found to be an optimal number of reduced texture features without compromising on classification accuracy. Scaled conjugate Gradient Back propagation network was used for classification using reduced texture features obtained from PCA analysis. By varying hidden layer neurons, accuracy of results achieved by proposed methods is analysed and is calculated to reach maximum accuracy with an optimal level of 15 neurons. Accuracy of 96.3% was achieved with 10 fractal features as input to neural network and 15 hidden layer neurons in neural network designed. The design of architecture is finalised with maximised accuracy for labelling microcalcification clusters as benign or malignant.

show abstract

Section: Automated Detection and Classification Of Microcalcificationmentioning

confidence: 92%

Automated Detection and Classification of Microcalcification Clusters with Enhanced Preprocessing and Fractal Analysis

Gowri

Valluvan²,

Chamundeeswari³

2018

Asian Pac J Cancer Prev

View full text Add to dashboard Cite

show abstract

“…Very similar to the flexible ReLU (FReLU) (see section 4.2.15) and dynamic ReLU (DReLU) (see section 4.2.14) is the displaced ReLU (DisReLU) 22 as it also shifts the ReLU [291]:…”

Section: Displaced Relu (Disrelu)mentioning

confidence: 99%

“…Since such notation would collide with the dynamic ReLU, we will use the original notation from [290] despite the inconsistency. 22 Macêdo et al originally abbreviated the displaced ReLU as DReLU but that is already taken by dynamic ReLU from section 4.2.14.…”

Section: Modified Lrelumentioning

confidence: 99%

“…Our aim is to provide researchers with a foundational resource that facilitates informed decision-making in selecting AFs for neural networks, recognizing that a more exhaustive exploration or detailed analysis would necessitate a dedicated and focused effort beyond the scope of this comprehensive listing. The presented overview is limited to real-valued activation functions; complex-valued neural networks (e.g., [3][4][5][6][7][8][9][10][11][12][13][14][15][16], brief overview available in [17,18]), bicomplex-valued neural networks (e.g., [19]), quaternion-valued neural networks (e.g., [20][21][22][23][24]), photonic neural networks (e.g., [25]), fuzzy neural networks (e.g., [26][27][28][29][30][31]), AFs for probabilistic boolean logic (e.g., [32]), quantum AFs (e.g., [33]) and others are out of the scope of this work. 1 We have chosen to categorize AFs into two main classes: fixed AFs (section 3) and adaptive activation functions (AAFs) (section 4), the latter having a parameter that is trained alongside the other weights in a network.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On Transformative Adaptive Activation Functions in Neural Networks for Gene Expression Inference

Kunc

Kléma²

2019

Preprint

View full text Add to dashboard Cite

Motivation: Gene expression profiling was made cheaper by the NIH LINCS program that profiles only ∼1, 000 selected landmark genes and uses them to reconstruct the whole profile. The D-GEX method employs neural networks to infer the whole profile. However, the original D-GEX can be further significantly improved. Results: We have analyzed the D-GEX method and determined that the inference can be improved using a logistic sigmoid activation function instead of the hyperbolic tangent. Moreover, we propose a novel transformative adaptive activation function that improves the gene expression inference even further and which generalizes several existing adaptive activation functions. Our improved neural network achieves average mean absolute error of 0.1340 which is a significant improvement over our reimplementation of the original D-GEX which achieves average mean absolute error 0.1637

show abstract

“…Without increasing the complexity of algorithm, the proposed RPROP algorithm is boosted by error-ependent weight backtracking step, which accelerates the training speed appreciably and provides better approximation accuracy. The neural network (ARENA et al 1996) (Minemoto et al 2016 and backpropagation algorithm in quaternionic domain (BP) (Cui, Takahashi, and Hashimoto 2013) has been widely applied in problems dealing with three and four dimensional information; recently its comparison with quaternionic scaled conjugate gradient (SCG) learning scheme is presented in (Popa 2016). This paper proposes an RPROP algorithm and compare with BP and SCG algorithms through application in 3D imaging and chaotic time series predictions.…”

Section: Resilient Algorithmmentioning

confidence: 99%

Untitled

2018

Preprint

View full text Add to dashboard Cite

This research uses Backpropagation Algorithm, Conjugate Gradient Fletcher-Reeves (CGFR) and Resilient. The purpose of this research is to see how much iteration and accuracy using this method compared with the level of iteration and accuracy in previous research using only backpropagation algorithm with Conjugate Gradient Fletcher-Reeves (CGFR) only in measuring consumer price index level. The data used as an example in this study is the Consumer Price Index (CPI) data based on foodstuffs sourced from the Central Statistics Agency Pematangsiantar Indonesia. There are 5 similar network architectures used in previous research and in this study for more objective results, including 12-6-1, 12-15-1, 12-24-1, 12-33-1 and 12-34-1. In the previous study, the best architecture was 12-15-1, with epoch level when using backpropagation algorithm of 821 iterations with 75% accuracy and Gradient fletcher reeves of 2 iterations with 67% accuracy. While the results of this study using the same architecture will be obtained epoch of 19 iterations with an accuracy of 50%. So it can be concluded that the use of backpropagation algorithm and gradient fletcher reeves to produce iteration and accuracy level better when compared with Resilient Algorithm.

show abstract

Scaled Conjugate Gradient Learning for Quaternion-Valued Neural Networks

Cited by 6 publications

References 21 publications

Automated Detection and Classification of Microcalcification Clusters with Enhanced Preprocessing and Fractal Analysis

Automated Detection and Classification of Microcalcification Clusters with Enhanced Preprocessing and Fractal Analysis

On Transformative Adaptive Activation Functions in Neural Networks for Gene Expression Inference

Untitled

Contact Info

Product

Resources

About