Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ IEEE Accessarrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
IEEE Access
Article . 2022 . Peer-reviewed
License: CC BY
Data sources: Crossref
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
IEEE Access
Article . 2022
Data sources: DOAJ
https://dx.doi.org/10.60692/t4...
Other literature type . 2022
Data sources: Datacite
https://dx.doi.org/10.60692/8j...
Other literature type . 2022
Data sources: Datacite
versions View all 4 versions
addClaim

This Research product is the result of merged Research products in OpenAIRE.

You have already added 0 works in your ORCID record related to the merged Research product.

An Interpretation of Long Short-Term Memory Recurrent Neural Network for Approximating Roots of Polynomials

تفسير للشبكة العصبية المتكررة للذاكرة طويلة المدى لتقريب جذور كثيرات الحدود
Authors: Madiha Bukhsh; Muhammad Saqib Ali; Muhammad Usman Ashraf; Khalid Alsubhi; Weiqiu Chen;

An Interpretation of Long Short-Term Memory Recurrent Neural Network for Approximating Roots of Polynomials

Abstract

Este artículo tiene como objetivo presentar un método flexible para interpretar la Red Neuronal Recurrente de Memoria a Largo Corto Plazo (LSTM-RNN) para la estructura relacional entre las raíces y los coeficientes de un polinomio. Primero se desarrolla una base de datos para entradas seleccionadas aleatoriamente en función de los grados del polinomio univariado que luego se utiliza para aproximar las raíces polinómicas a través del modelo LSTM-RNN propuesto. Además, se utiliza un algoritmo de optimización de aprendizaje adaptativo específicamente para actualizar los pesos de la red de forma iterativa basándose en el entrenamiento de datos de redes neuronales profundas. Por lo tanto, el método puede explotar la capacidad de encontrar las tasas de aprendizaje individuales para cada variable a través de estrategias de tasa de aprendizaje adaptativo para evitar de manera efectiva que los pesos fluctúen en un amplio espectro. Finalmente, se realizan varios resultados experimentales que muestran que el modelo LSTM-RNN propuesto se puede utilizar como un enfoque alternativo para calcular una aproximación de cada raíz para un polinomio dado. Además, los resultados se comparan con el modelo de red neuronal artificial basado en redes neuronales feedforward convencionales. Los resultados demuestran claramente la superioridad del modelo LSTM-RNN propuesto para la aproximación de raíces en términos de precisión, error cuadrático medio y convergencia más rápida.

Cet article vise à présenter une méthode flexible d'interprétation du Long Short-Term Memory Recurrent Neural Network (LSTM-RNN) pour la structure relationnelle entre les racines et les coefficients d'un polynôme. Une base de données est d'abord développée pour les entrées sélectionnées aléatoirement en fonction des degrés du polynôme univarié qui est ensuite utilisé pour approximer les racines polynomiales à travers le modèle LSTM-RNN proposé. En outre, un algorithme d'optimisation d'apprentissage adaptatif est utilisé spécifiquement pour mettre à jour les poids de réseau de manière itérative sur la base de données de réseaux neuronaux profonds d'entraînement. Ainsi, le procédé peut exploiter la capacité de trouver les taux d'apprentissage individuels pour chaque variable grâce à des stratégies de taux d'apprentissage adaptatives pour empêcher efficacement les poids de fluctuer dans un large spectre. Enfin, plusieurs résultats expérimentaux sont réalisés qui montrent que le modèle LSTM-RNN proposé peut être utilisé comme une approche alternative pour calculer une approximation de chaque racine pour un polynôme donné. En outre, les résultats sont comparés au modèle de réseau neuronal artificiel conventionnel basé sur un réseau neuronal à anticipation. Les résultats démontrent clairement la supériorité du modèle LSTM-RNN proposé pour l'approximation des racines en termes de précision, d'erreur quadratique moyenne et de convergence plus rapide.

This paper aims to present a flexible method for interpreting the Long Short-Term Memory Recurrent Neural Network (LSTM-RNN) for the relational structure between the roots and the coefficients of a polynomial. A database is first developed for randomly selected inputs based on the degrees of the univariate polynomial which is then used to approximate the polynomial roots through the proposed LSTM-RNN model. Furthermore, an adaptive learning optimization algorithm is used specifically to update the network weights iteratively based on training deep neural networks data. Thus, the method can exploit the ability to find the individual learning rates for each variable through adaptive learning rate strategies to effectively prevent the weights from fluctuating in a wide spectrum. Finally, several experimental results are performed which shows that the proposed LSTM-RNN model can be used as an alternative approach to compute an approximation of each root for a given polynomial. Furthermore, the results are compared with the conventional feedforward neural network based artificial neural network model. The results clearly demonstrate the superiority of the proposed LSTM-RNN model for roots approximation in terms of accuracy, mean square error and faster convergence.

تهدف هذه الورقة إلى تقديم طريقة مرنة لتفسير الشبكة العصبية المتكررة للذاكرة طويلة المدى (LSTM - RNN) للبنية العلائقية بين جذور ومعاملات متعدد الحدود. يتم تطوير قاعدة بيانات أولاً للمدخلات المختارة عشوائيًا بناءً على درجات متعدد الحدود أحادي المتغير والذي يتم استخدامه بعد ذلك لتقريب جذور متعدد الحدود من خلال نموذج LSTM - RNN المقترح. علاوة على ذلك، يتم استخدام خوارزمية تحسين التعلم التكيفي على وجه التحديد لتحديث أوزان الشبكة بشكل متكرر بناءً على تدريب بيانات الشبكات العصبية العميقة. وبالتالي، يمكن للطريقة استغلال القدرة على العثور على معدلات التعلم الفردية لكل متغير من خلال استراتيجيات معدل التعلم التكيفي لمنع الأوزان بشكل فعال من التقلب في طيف واسع. أخيرًا، يتم إجراء العديد من النتائج التجريبية التي توضح أنه يمكن استخدام نموذج LSTM - RNN المقترح كنهج بديل لحساب تقريب كل جذر لكثير حدود معين. علاوة على ذلك، تتم مقارنة النتائج مع نموذج الشبكة العصبية الاصطناعية التقليدية القائمة على التغذية الأمامية. تُظهر النتائج بوضوح تفوق نموذج LSTM - RNN المقترح لتقريب الجذور من حيث الدقة ومتوسط الخطأ المربع والتقارب الأسرع.

Keywords

Function Approximation, Backpropagation Learning, Artificial neural network, Artificial intelligence, Economics, Recurrent neural network, Polynomial, Mathematical analysis, Artificial Intelligence, Machine learning, Long short-term memory, Univariate, FOS: Mathematics, Model Reduction, Floating-Point Arithmetic in Scientific Computation, Recurrent Neural Networks, Interval Analysis, Economic growth, Feedforward neural network, Physics-Informed Neural Networks for Scientific Computing, Statistics, deep neural network, Statistical and Nonlinear Physics, Neural Network Fundamentals and Applications, Computer science, Multivariate statistics, TK1-9971, Algorithm, Computational Theory and Mathematics, Physics and Astronomy, adaptive moment estimation algorithm, Computer Science, Physical Sciences, Convergence (economics), Mean squared error, recurrent neural network, Electrical engineering. Electronics. Nuclear engineering, error cost function, Mathematics

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    4
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Top 10%
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
4
Top 10%
Average
Average
gold