Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability

The current systematic logical rules in the Discrete Hopfield Neural Network encounter significant challenges, including repetitive final neuron states that lead to the issue of overfitting. Furthermore, the systematic logical rules neglect the impact on the appearance of negative literals within th...

Full description

Bibliographic Details
Main Authors: Rusdi, Nur ‘Afifah, Zamri, Nur Ezlin, Kasihmuddin, Mohd Shareduwan Mohd, Romli, Nurul Atiqah, Manoharam, Gaeithry, Abdeen, Suad, Mansor, Mohd. Asyraf
Format: Article
Language:English
Published: American Institute of Mathematical Sciences (AIMS) 2024
Online Access:http://psasir.upm.edu.my/id/eprint/116164/
http://psasir.upm.edu.my/id/eprint/116164/1/116164.pdf
_version_ 1848866938666614784
author Rusdi, Nur ‘Afifah
Zamri, Nur Ezlin
Kasihmuddin, Mohd Shareduwan Mohd
Romli, Nurul Atiqah
Manoharam, Gaeithry
Abdeen, Suad
Mansor, Mohd. Asyraf
author_facet Rusdi, Nur ‘Afifah
Zamri, Nur Ezlin
Kasihmuddin, Mohd Shareduwan Mohd
Romli, Nurul Atiqah
Manoharam, Gaeithry
Abdeen, Suad
Mansor, Mohd. Asyraf
author_sort Rusdi, Nur ‘Afifah
building UPM Institutional Repository
collection Online Access
description The current systematic logical rules in the Discrete Hopfield Neural Network encounter significant challenges, including repetitive final neuron states that lead to the issue of overfitting. Furthermore, the systematic logical rules neglect the impact on the appearance of negative literals within the logical structure, and most recent efforts have primarily focused on improving the learning capabilities of the network, which could potentially limit its overall efficiency. To tackle the limitation, we introduced a Negative Based Higher Order Systematic Logic to the network, imposing restriction on the appearance of negative literals within the clauses. Additionally, a Hybrid Black Hole Algorithm was proposed in the retrieval phase to optimize the final neuron states. This ensured that the optimized states achieved maximum diversity and reach global minima solutions with the lowest similarity index, thereby enhancing the overall performance of the network. The results illustrated that the proposed model can achieve up to 10,000 diversified and global solutions with an average similarity index of 0.09. The findings indicated that the optimized final neuron states are in optimal configurations. Based on the findings, the development of the new systematic SAT and the implementation of the Hybrid Black Hole algorithm to optimize the retrieval capabilities of DHNN to achieve multi-objective functions result in updated final neuron states with high diversity, high attainment of global minima solutions, and produces states with a low similarity index. Consequently, this proposed model could be extended for logic mining applications to tackle classification tasks. The optimized final neuron states will enhance the retrieval of high-quality induced logic, which is effective for classification and knowledge extraction.
first_indexed 2025-11-15T14:28:33Z
format Article
id upm-116164
institution Universiti Putra Malaysia
institution_category Local University
language English
last_indexed 2025-11-15T14:28:33Z
publishDate 2024
publisher American Institute of Mathematical Sciences (AIMS)
recordtype eprints
repository_type Digital Repository
spelling upm-1161642025-03-19T08:16:44Z http://psasir.upm.edu.my/id/eprint/116164/ Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability Rusdi, Nur ‘Afifah Zamri, Nur Ezlin Kasihmuddin, Mohd Shareduwan Mohd Romli, Nurul Atiqah Manoharam, Gaeithry Abdeen, Suad Mansor, Mohd. Asyraf The current systematic logical rules in the Discrete Hopfield Neural Network encounter significant challenges, including repetitive final neuron states that lead to the issue of overfitting. Furthermore, the systematic logical rules neglect the impact on the appearance of negative literals within the logical structure, and most recent efforts have primarily focused on improving the learning capabilities of the network, which could potentially limit its overall efficiency. To tackle the limitation, we introduced a Negative Based Higher Order Systematic Logic to the network, imposing restriction on the appearance of negative literals within the clauses. Additionally, a Hybrid Black Hole Algorithm was proposed in the retrieval phase to optimize the final neuron states. This ensured that the optimized states achieved maximum diversity and reach global minima solutions with the lowest similarity index, thereby enhancing the overall performance of the network. The results illustrated that the proposed model can achieve up to 10,000 diversified and global solutions with an average similarity index of 0.09. The findings indicated that the optimized final neuron states are in optimal configurations. Based on the findings, the development of the new systematic SAT and the implementation of the Hybrid Black Hole algorithm to optimize the retrieval capabilities of DHNN to achieve multi-objective functions result in updated final neuron states with high diversity, high attainment of global minima solutions, and produces states with a low similarity index. Consequently, this proposed model could be extended for logic mining applications to tackle classification tasks. The optimized final neuron states will enhance the retrieval of high-quality induced logic, which is effective for classification and knowledge extraction. American Institute of Mathematical Sciences (AIMS) 2024 Article PeerReviewed text en cc_by_4 http://psasir.upm.edu.my/id/eprint/116164/1/116164.pdf Rusdi, Nur ‘Afifah and Zamri, Nur Ezlin and Kasihmuddin, Mohd Shareduwan Mohd and Romli, Nurul Atiqah and Manoharam, Gaeithry and Abdeen, Suad and Mansor, Mohd. Asyraf (2024) Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability. AIMS Mathematics, 9 (11). pp. 29820-29882. ISSN 2473-6988 http://www.aimspress.com/article/doi/10.3934/math.20241444 10.3934/math.20241444
spellingShingle Rusdi, Nur ‘Afifah
Zamri, Nur Ezlin
Kasihmuddin, Mohd Shareduwan Mohd
Romli, Nurul Atiqah
Manoharam, Gaeithry
Abdeen, Suad
Mansor, Mohd. Asyraf
Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability
title Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability
title_full Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability
title_fullStr Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability
title_full_unstemmed Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability
title_short Synergizing intelligence and knowledge discovery: Hybrid black hole algorithm for optimizing discrete Hopfield neural network with negative based systematic satisfiability
title_sort synergizing intelligence and knowledge discovery: hybrid black hole algorithm for optimizing discrete hopfield neural network with negative based systematic satisfiability
url http://psasir.upm.edu.my/id/eprint/116164/
http://psasir.upm.edu.my/id/eprint/116164/
http://psasir.upm.edu.my/id/eprint/116164/
http://psasir.upm.edu.my/id/eprint/116164/1/116164.pdf