Mathew Mithra Noel
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Mathew Mithra Noel has a total of 39 co-authors. Here are the top co-authors by number of shared publications:
- D. Geraldine Bessie Amali
Vellore Institute of Technology
5 shared publications - Venkataraman Muthiah-Nakarajan
School of Electrical Engineering Vellore Institute of Technology
5 shared publications - Balakrishnan Jaganatha Pandian
School of Electrical Engineering, VIT University, Vellore - 632014, Tamil Nadu, India
4 shared publications T.C. Jannett
4 shared publications- Ankit Sharma
Department of Electrical and Computer Engineering, Gannon University, Erie, PA 16541, USA
3 shared publications - Ruban Nersisson
School of Electrical Engineering, Vellore Institute of Technology, Vellore, India
3 shared publications
Latest Publications
- Open Access
- 3 August 2026
- Published by Elsevier in SSRN Electronic Journal
Alternate loss functions and regression models that achieve robustness to outliers by modulating the learning rate
Most real-world datasets used for training supervised learning models are contaminated with noisy data and outliers that complicate the model parameter learning process. This paper proposes a new approach for achieving robustness where the learning rate is modulated by a factor that is sensitive to outliers. In this approach a reduction of the learning rate is shown to be achieved by using alternate loss functions that are infinitely differentiable, strictly convex or quasiconvex and more closely approximate the absolute error than Huber and log-cosh losses. A comparison of the performance of regression models trained with different loss functions on a wide variety of benchmarks and datasets is presented to demonstrate the superior performance of the Square Root Loss (SRL) and Smooth Mean Absolute Error (SMAE) losses proposed in this paper. Two new robust linear regression models are presented. Highly vectorized robust parameter update formulae that take advantage of modern GPUs for both stochastic and batch gradient descent are presented. Finally, a new robust gradient descent parameter update formula for nonlinear regression models is presented.
- 3 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessAlternate loss functions and regression models that achieve robustness to outliers by modulating the learning rate
Most real-world datasets used for training supervised learning models are contaminated with noisy data and outliers that complicate the model parameter learning process. This paper proposes a new approach for achieving robustness where the learning rate is modulated by a factor that is sensitive to outliers. In this approach a reduction of the learning rate is shown to be achieved by using alternate loss functions that are infinitely differentiable, strictly convex or quasiconvex and more closely approximate the absolute error than Huber and log-cosh losses. A comparison of the performance of regression models trained with different loss functions on a wide variety of benchmarks and datasets is presented to demonstrate the superior performance of the Square Root Loss (SRL) and Smooth Mean Absolute Error (SMAE) losses proposed in this paper. Two new robust linear regression models are presented. Highly vectorized robust parameter update formulae that take advantage of modern GPUs for both stochastic and batch gradient descent are presented. Finally, a new robust gradient descent parameter update formula for nonlinear regression models is presented.
- Open Access
- 17 April 2026
- Published by Elsevier in SSRN Electronic Journal
Alternate Loss Functions for Classification and Robust Regression Can Improve the Accuracy of Artificial Neural Networks
All machine learning algorithms use a loss, cost, utility, or reward function to encode the learning objective and oversee the learning process. This paper shows that the training speed, final accuracy and robustness to outliers can significantly depend on the loss function used to train neural networks. In particular derivative of the loss with respect to the output can be significantly different with different loss functions leading to significantly different performance after gradient-descent based training. This paper explores the possible advantages of using new loss functions that penalize errors differently compared to the popular Cross-entropy and Mean squared loss on 7 benchmarks. Two new loss functions for classification are proposed and switching loss functions during training is shown to significantly speedup training. Two new infinitely differentiable loss function that more closely approximate the absolute error loss compared to the Huber and Log-Cosh losses used for robust regression are proposed and shown to significantly outperform Squared error, Huber, and Log-Cosh losses on datasets contaminated with outliers.
- 17 April 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessAlternate Loss Functions for Classification and Robust Regression Can Improve the Accuracy of Artificial Neural Networks
All machine learning algorithms use a loss, cost, utility, or reward function to encode the learning objective and oversee the learning process. This paper shows that the training speed, final accuracy and robustness to outliers can significantly depend on the loss function used to train neural networks. In particular derivative of the loss with respect to the output can be significantly different with different loss functions leading to significantly different performance after gradient-descent based training. This paper explores the possible advantages of using new loss functions that penalize errors differently compared to the popular Cross-entropy and Mean squared loss on 7 benchmarks. Two new loss functions for classification are proposed and switching loss functions during training is shown to significantly speedup training. Two new infinitely differentiable loss function that more closely approximate the absolute error loss compared to the Huber and Log-Cosh losses used for robust regression are proposed and shown to significantly outperform Squared error, Huber, and Log-Cosh losses on datasets contaminated with outliers.
- Open Access
- 11 August 2025
- Published by MDPI AG in Quantum Reports
- Vol. 7 (3) , 36
Quantum-Enhanced Algorithmic Fairness and the Advancement of AI Integrity and Responsibility
In the evolving digital landscape, the pervasive influence of artificial intelligence (AI) on social media platforms reveals a compelling paradox: the capability to provide personalized experiences juxtaposed with inherent biases reminiscent of human imperfections. Such biases prompt rigorous contemplation on matters of fairness, equity, and societal ramifications, and penetrate the foundational essence of AI. Within this intricate context, the present work ventures into novel domains by examining the potential of quantum computing as a viable remedy for bias in artificial intelligence. The conceptual framework of the quantum sentinel is presented—an innovative approach that employs quantum principles for the detection and scrutiny of biases in AI algorithms. Furthermore, the study poses and investigates the question of whether the integration of advanced quantum computing to address AI bias is seen as an excessive measure or a requisite advancement commensurate with the intricacy of the issue. By intertwining quantum mechanics, AI bias, and the philosophical considerations they induce, this research fosters a discourse on the journey toward ethical AI, thus establishing a foundation for an ethically conscious and balanced digital environment. We also show that the quantum Zeno effect can protect SVM hyperplanes from bias through targeted simulations.
- 11 August 2025
- Published by MDPI AG in Quantum Reports
- Vol. 7 (3) , 36
Open AccessQuantum-Enhanced Algorithmic Fairness and the Advancement of AI Integrity and Responsibility
In the evolving digital landscape, the pervasive influence of artificial intelligence (AI) on social media platforms reveals a compelling paradox: the capability to provide personalized experiences juxtaposed with inherent biases reminiscent of human imperfections. Such biases prompt rigorous contemplation on matters of fairness, equity, and societal ramifications, and penetrate the foundational essence of AI. Within this intricate context, the present work ventures into novel domains by examining the potential of quantum computing as a viable remedy for bias in artificial intelligence. The conceptual framework of the quantum sentinel is presented—an innovative approach that employs quantum principles for the detection and scrutiny of biases in AI algorithms. Furthermore, the study poses and investigates the question of whether the integration of advanced quantum computing to address AI bias is seen as an excessive measure or a requisite advancement commensurate with the intricacy of the issue. By intertwining quantum mechanics, AI bias, and the philosophical considerations they induce, this research fosters a discourse on the journey toward ethical AI, thus establishing a foundation for an ethically conscious and balanced digital environment. We also show that the quantum Zeno effect can protect SVM hyperplanes from bias through targeted simulations.
- 11 April 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 175
Oscillating activation functions can improve the performance of convolutional neural networks
- 11 April 2025
- Published by Elsevier in Applied Soft Computing
- Vol. 175
Oscillating activation functions can improve the performance of convolutional neural networks
- 13 December 2024
- Published by Elsevier in Expert Systems with Applications
- Vol. 266
Biologically inspired oscillating activation functions can bridge the performance gap between biological and artificial neurons
- 13 December 2024
- Published by Elsevier in Expert Systems with Applications
- Vol. 266
Biologically inspired oscillating activation functions can bridge the performance gap between biological and artificial neurons
- Open Access
- 23 September 2024
- Published by MDPI AG in Preprints.org
Quantum-Enhanced Algorithmic Fairness and the Advancement of AI Integrity and Responsibility
In the evolving digital landscape, the pervasive influence of Artificial Intelligence (AI) on social media platforms reveals a compelling paradox: the capability to provide personalized experiences juxtaposed with inherent biases reminiscent of human imperfections. Such biases prompt rigorous contemplation on matters of fairness, equity, and societal ramifications, and penetrate the foundational essence of AI. Within this intricate context, the research ventures into novel domains by examining the potential of quantum computing as a viable remedy for bias in artificial intelligence. The conceptual framework of the Quantum Sentinel is presented - an innovative approach that employs quantum principles for the detection and scrutiny of biases in AI algorithms. Furthermore, the study poses and investigates the question of whether the integration of advanced quantum computing to address AI bias is seen as an excessive measure or a requisite advancement commensurate with the intricacy of the issue. By intertwining quantum mechanics, AI bias, and the philosophical considerations they induce, this research fosters a discourse on the journey toward ethical AI, thus establishing a foundation for an ethically conscious and balanced digital environment.
- 23 September 2024
- Published by MDPI AG in Preprints.org
Open AccessQuantum-Enhanced Algorithmic Fairness and the Advancement of AI Integrity and Responsibility
In the evolving digital landscape, the pervasive influence of Artificial Intelligence (AI) on social media platforms reveals a compelling paradox: the capability to provide personalized experiences juxtaposed with inherent biases reminiscent of human imperfections. Such biases prompt rigorous contemplation on matters of fairness, equity, and societal ramifications, and penetrate the foundational essence of AI. Within this intricate context, the research ventures into novel domains by examining the potential of quantum computing as a viable remedy for bias in artificial intelligence. The conceptual framework of the Quantum Sentinel is presented - an innovative approach that employs quantum principles for the detection and scrutiny of biases in AI algorithms. Furthermore, the study poses and investigates the question of whether the integration of advanced quantum computing to address AI bias is seen as an excessive measure or a requisite advancement commensurate with the intricacy of the issue. By intertwining quantum mechanics, AI bias, and the philosophical considerations they induce, this research fosters a discourse on the journey toward ethical AI, thus establishing a foundation for an ethically conscious and balanced digital environment.
- Open Access
A Significantly Better Class of Activation Functions Than ReLU Like Activation Functions
This paper introduces a significantly better class of activation functions than the almost universally used ReLU like and Sigmoidal class of activation functions. Two new activation functions referred to as the Cone and Parabolic-Cone that differ drastically from popular activation functions and significantly outperform these on the CIFAR-10 and Imagenette benchmmarks are proposed. The cone activation functions are positive only on a finite interval and are strictly negative except at the end-points of the interval, where they become zero. Thus the set of inputs that produce a positive output for a neuron with cone activation functions is a hyperstrip and not a half-space as is the usual case. Since a hyper strip is the region between two parallel hyper-planes, it allows neurons to more finely divide the input feature space into positive and negative classes than with infinitely wide half-spaces. In particular the XOR function can be learn by a single neuron with cone-like activation functions. Both the cone and parabolic-cone activation functions are shown to achieve higher accuracies with significantly fewer neurons on benchmarks. The results presented in this paper indicate that many nonlinear real-world datasets may be separated with fewer hyperstrips than half-spaces. The Cone and Parabolic-Cone activation functions have larger derivatives than ReLU and are shown to significantly speedup training.
Open AccessA Significantly Better Class of Activation Functions Than ReLU Like Activation Functions
This paper introduces a significantly better class of activation functions than the almost universally used ReLU like and Sigmoidal class of activation functions. Two new activation functions referred to as the Cone and Parabolic-Cone that differ drastically from popular activation functions and significantly outperform these on the CIFAR-10 and Imagenette benchmmarks are proposed. The cone activation functions are positive only on a finite interval and are strictly negative except at the end-points of the interval, where they become zero. Thus the set of inputs that produce a positive output for a neuron with cone activation functions is a hyperstrip and not a half-space as is the usual case. Since a hyper strip is the region between two parallel hyper-planes, it allows neurons to more finely divide the input feature space into positive and negative classes than with infinitely wide half-spaces. In particular the XOR function can be learn by a single neuron with cone-like activation functions. Both the cone and parabolic-cone activation functions are shown to achieve higher accuracies with significantly fewer neurons on benchmarks. The results presented in this paper indicate that many nonlinear real-world datasets may be separated with fewer hyperstrips than half-spaces. The Cone and Parabolic-Cone activation functions have larger derivatives than ReLU and are shown to significantly speedup training.
- Open Access
Growing Cosine Unit: A Novel Oscillatory Activation Function That Can Speedup Training and Reduce Parameters in Convolutional Neural Networks
Convolutional neural networks have been successful in solving many socially important and economically significant problems. This ability to learn complex high-dimensional functions hierarchically can be attributed to the use of nonlinear activation functions. A key discovery that made training deep networks feasible was the adoption of the Rectified Linear Unit (ReLU) activation function to alleviate the vanishing gradient problem caused by using saturating activation functions. Since then, many improved variants of the ReLU activation have been proposed. However, a majority of activation functions used today are non-oscillatory and monotonically increasing due to their biological plausibility. This paper demonstrates that oscillatory activation functions can improve gradient flow and reduce network size. Two theorems on limits of non-oscillatory activation functions are presented. A new oscillatory activation function called Growing Cosine Unit(GCU) defined as $C(z) = z\cos z$ that outperforms Sigmoids, Swish, Mish and ReLU on a variety of architectures and benchmarks is presented. The GCU activation has multiple zeros enabling single GCU neurons to have multiple hyperplanes in the decision boundary. This allows single GCU neurons to learn the XOR function without feature engineering. Experimental results indicate that replacing the activation function in the convolution layers with the GCU activation function significantly improves performance on CIFAR-10, CIFAR-100 and Imagenette.Open AccessGrowing Cosine Unit: A Novel Oscillatory Activation Function That Can Speedup Training and Reduce Parameters in Convolutional Neural Networks
Convolutional neural networks have been successful in solving many socially important and economically significant problems. This ability to learn complex high-dimensional functions hierarchically can be attributed to the use of nonlinear activation functions. A key discovery that made training deep networks feasible was the adoption of the Rectified Linear Unit (ReLU) activation function to alleviate the vanishing gradient problem caused by using saturating activation functions. Since then, many improved variants of the ReLU activation have been proposed. However, a majority of activation functions used today are non-oscillatory and monotonically increasing due to their biological plausibility. This paper demonstrates that oscillatory activation functions can improve gradient flow and reduce network size. Two theorems on limits of non-oscillatory activation functions are presented. A new oscillatory activation function called Growing Cosine Unit(GCU) defined as $C(z) = z\cos z$ that outperforms Sigmoids, Swish, Mish and ReLU on a variety of architectures and benchmarks is presented. The GCU activation has multiple zeros enabling single GCU neurons to have multiple hyperplanes in the decision boundary. This allows single GCU neurons to learn the XOR function without feature engineering. Experimental results indicate that replacing the activation function in the convolution layers with the GCU activation function significantly improves performance on CIFAR-10, CIFAR-100 and Imagenette. - 9 December 2022
- Published by IGI Global Scientific Publishing
- p. 132-155
Automatic Image Captioning Using Different Variants of the Long Short-Term Memory (LSTM) Deep Learning Model
Today's world is full of digital images; however, the context is unavailable most of the time. Thus, image captioning is quintessential for providing the content of an image. Besides generating accurate captions, the image captioning model must also be scalable. In this chapter, two variants of long short-term memory (LSTM), namely stacked LSTM and BiLSTM along with convolutional neural networks (CNN) have been used to implement the Encoder-Decoder model for generating captions. Bilingual evaluation understudy (BLEU) score metric is used to evaluate the performance of these two bi-layered models. From the study, it was observed that both the models were on par when it came to performance. Some resulted in low BLEU scores suggesting that the predicted caption was dissimilar to the actual caption whereas some very high BLEU scores suggested that the model was able to predict captions almost similar to human. Furthermore, it was found that the bidirectional LSTM model is more computationally intensive and requires more time to train than the stacked LSTM model owing to its complex architecture.
- 9 December 2022
- Published by IGI Global Scientific Publishing
- p. 132-155
Automatic Image Captioning Using Different Variants of the Long Short-Term Memory (LSTM) Deep Learning Model
Today's world is full of digital images; however, the context is unavailable most of the time. Thus, image captioning is quintessential for providing the content of an image. Besides generating accurate captions, the image captioning model must also be scalable. In this chapter, two variants of long short-term memory (LSTM), namely stacked LSTM and BiLSTM along with convolutional neural networks (CNN) have been used to implement the Encoder-Decoder model for generating captions. Bilingual evaluation understudy (BLEU) score metric is used to evaluate the performance of these two bi-layered models. From the study, it was observed that both the models were on par when it came to performance. Some resulted in low BLEU scores suggesting that the predicted caption was dissimilar to the actual caption whereas some very high BLEU scores suggested that the model was able to predict captions almost similar to human. Furthermore, it was found that the bidirectional LSTM model is more computationally intensive and requires more time to train than the stacked LSTM model owing to its complex architecture.
- 1 November 2020
- Published by Elsevier in Applied Soft Computing
- Vol. 96, 106709
Control of constrained high dimensional nonlinear liquid level processes using a novel neural network based Rapidly exploring Random Tree algorithm
- 1 November 2020
- Published by Elsevier in Applied Soft Computing
- Vol. 96, 106709
Control of constrained high dimensional nonlinear liquid level processes using a novel neural network based Rapidly exploring Random Tree algorithm