Ashish Kumar Pandey
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
Ashish Kumar Pandey has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Koushik Biswas
Northwestern University
7 shared publications - Sandeep Kumar
IIIT Delhi
6 shared publications - Atul Kumar
Department of Mathematics, IIIT, Delhi, India
2 shared publications Bhavna
2 shared publications- Shilpak Banerjee
Department of Mathematics and Statistics Indian Institute of Technology Tirupati 517619 Tirupati India
2 shared publications - Shilpak Banerjee
Department of Mathematics IIIT Delhi 110020 New Delhi India
2 shared publications
Latest Publications
- Open Access
Spectral stability of periodic traveling waves in Caudrey-Dodd-Gibbon-Sawada-Kotera Equation
We study the spectral stability of the one-dimensional small-amplitude periodic traveling wave solutions of the (1+1)-dimensional Caudrey-Dodd-Gibbon-Sawada-Kotera equation. We show that these waves are spectrally stable with respect to co-periodic as well as square integrable perturbations.
Open AccessSpectral stability of periodic traveling waves in Caudrey-Dodd-Gibbon-Sawada-Kotera Equation
We study the spectral stability of the one-dimensional small-amplitude periodic traveling wave solutions of the (1+1)-dimensional Caudrey-Dodd-Gibbon-Sawada-Kotera equation. We show that these waves are spectrally stable with respect to co-periodic as well as square integrable perturbations.
- Open Access
Forecasting formation of a Tropical Cyclone Using Reanalysis Data
The tropical cyclone formation process is one of the most complex natural phenomena which is governed by various atmospheric, oceanographic, and geographic factors that varies with time and space. Despite several years of research, accurately predicting tropical cyclone formation remains a challenging task. While the existing numerical models have inherent limitations, the machine learning models fail to capture the spatial and temporal dimensions of the causal factors behind TC formation. In this study, a deep learning model has been proposed that can forecast the formation of a tropical cyclone with a lead time of up to 60 hours with high accuracy. The model uses the high-resolution reanalysis data ERA5 (ECMWF reanalysis 5th generation), and best track data IBTrACS (International Best Track Archive for Climate Stewardship) to forecast tropical cyclone formation in six ocean basins of the world. For 60 hours lead time the models achieve an accuracy in the range of 86.9% - 92.9% across the six ocean basins. The model takes about 5-15 minutes of training time depending on the ocean basin, and the amount of data used and can predict within seconds, thereby making it suitable for real-life usage.
Open AccessForecasting formation of a Tropical Cyclone Using Reanalysis Data
The tropical cyclone formation process is one of the most complex natural phenomena which is governed by various atmospheric, oceanographic, and geographic factors that varies with time and space. Despite several years of research, accurately predicting tropical cyclone formation remains a challenging task. While the existing numerical models have inherent limitations, the machine learning models fail to capture the spatial and temporal dimensions of the causal factors behind TC formation. In this study, a deep learning model has been proposed that can forecast the formation of a tropical cyclone with a lead time of up to 60 hours with high accuracy. The model uses the high-resolution reanalysis data ERA5 (ECMWF reanalysis 5th generation), and best track data IBTrACS (International Best Track Archive for Climate Stewardship) to forecast tropical cyclone formation in six ocean basins of the world. For 60 hours lead time the models achieve an accuracy in the range of 86.9% - 92.9% across the six ocean basins. The model takes about 5-15 minutes of training time depending on the ocean basin, and the amount of data used and can predict within seconds, thereby making it suitable for real-life usage.
- 7 November 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (8) , 5807-5818
Will a tropical cyclone make landfall?
- 7 November 2022
- Published by Springer Nature in Neural Computing and Applications
- Vol. 35 (8) , 5807-5818
Will a tropical cyclone make landfall?
- 23 October 2022
- Published by Springer Nature
- p. 313-329
SAU: Smooth Activation Function Using Convolution with Approximate Identities
- 23 October 2022
- Published by Springer Nature
- p. 313-329
SAU: Smooth Activation Function Using Convolution with Approximate Identities
- Open Access
- 13 April 2022
- Published by The Royal Society in Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences
- Vol. 478 (2260)
Transverse spectral instability in generalized Kadomtsev–Petviashvili equation
We study transverse stability and instability of one-dimensional small-amplitude periodic travelling waves of a generalized Kadomtsev–Petviashvili equation with respect to two-dimensional perturbations, which are either periodic or square-integrable in the direction of the propagation of the underlying one-dimensional wave and periodic in the transverse direction. We obtain transverse instability results in KP-fKdV, KP-ILW and KP-Whitham equations. Moreover, assuming the spectral stability of one-dimensional wave with respect to one-dimensional square-integrable periodic perturbations, we obtain transverse stability results in the aforementioned equations.
- 13 April 2022
- Published by The Royal Society in Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences
- Vol. 478 (2260)
Open AccessTransverse spectral instability in generalized Kadomtsev–Petviashvili equation
We study transverse stability and instability of one-dimensional small-amplitude periodic travelling waves of a generalized Kadomtsev–Petviashvili equation with respect to two-dimensional perturbations, which are either periodic or square-integrable in the direction of the propagation of the underlying one-dimensional wave and periodic in the transverse direction. We obtain transverse instability results in KP-fKdV, KP-ILW and KP-Whitham equations. Moreover, assuming the spectral stability of one-dimensional wave with respect to one-dimensional square-integrable periodic perturbations, we obtain transverse stability results in the aforementioned equations.
- Open Access
Transverse Spectral Instabilities in Konopelchenko-Dubrovsky Equation
We study the transverse spectral stability of the one-dimensional small-amplitude periodic traveling wave solutions of the (2+1)-dimensional Konopelchenko-Dubrovsky (KD) equation. We show that these waves are transversely unstable with respect to two-dimensional perturbations that are periodic in both directions with long wavelength in the transverse direction. We also show that these waves are transversely stable with respect to perturbations which are either mean-zero periodic or square-integrable in the direction of the propagation of the wave and periodic in the transverse direction with finite or short wavelength. We discuss the implications of these results for special cases of the KD equation - namely, KP-II and mKP-II equations.
Open AccessTransverse Spectral Instabilities in Konopelchenko-Dubrovsky Equation
We study the transverse spectral stability of the one-dimensional small-amplitude periodic traveling wave solutions of the (2+1)-dimensional Konopelchenko-Dubrovsky (KD) equation. We show that these waves are transversely unstable with respect to two-dimensional perturbations that are periodic in both directions with long wavelength in the transverse direction. We also show that these waves are transversely stable with respect to perturbations which are either mean-zero periodic or square-integrable in the direction of the propagation of the wave and periodic in the transverse direction with finite or short wavelength. We discuss the implications of these results for special cases of the KD equation - namely, KP-II and mKP-II equations.
- Open Access
Transverse spectral instability in generalized Kadomtsev-Petviashvili equation
We study transverse stability and instability of one-dimensional small-amplitude periodic traveling waves of a generalized Kadomtsev-Petviashvili equation with respect to two-dimensional perturbations, which are either periodic or square-integrable in the direction of the propagation of the underlying one-dimensional wave and periodic in the transverse direction. We obtain transverse instability results in KP-fKdV, KP-ILW, and KP-Whitham equations. Moreover, assuming the spectral stability of one-dimensional wave with respect to one-dimensional square-integrable periodic perturbations, we obtain transverse stability results in aforementioned equations.
Open AccessTransverse spectral instability in generalized Kadomtsev-Petviashvili equation
We study transverse stability and instability of one-dimensional small-amplitude periodic traveling waves of a generalized Kadomtsev-Petviashvili equation with respect to two-dimensional perturbations, which are either periodic or square-integrable in the direction of the propagation of the underlying one-dimensional wave and periodic in the transverse direction. We obtain transverse instability results in KP-fKdV, KP-ILW, and KP-Whitham equations. Moreover, assuming the spectral stability of one-dimensional wave with respect to one-dimensional square-integrable periodic perturbations, we obtain transverse stability results in aforementioned equations.
- Open Access
ErfAct and Pserf: Non-monotonic Smooth Trainable Activation Functions
An activation function is a crucial component of a neural network that introduces non-linearity in the network. The state-of-the-art performance of a neural network depends also on the perfect choice of an activation function. We propose two novel non-monotonic smooth trainable activation functions, called ErfAct and Pserf. Experiments suggest that the proposed functions improve the network performance significantly compared to the widely used activations like ReLU, Swish, and Mish. Replacing ReLU by ErfAct and Pserf, we have 5.68% and 5.42% improvement for top-1 accuracy on Shufflenet V2 (2.0x) network in CIFAR100 dataset, 2.11% and 1.96% improvement for top-1 accuracy on Shufflenet V2 (2.0x) network in CIFAR10 dataset, 1.0%, and 1.0% improvement on mean average precision (mAP) on SSD300 model in Pascal VOC dataset.
Open AccessErfAct and Pserf: Non-monotonic Smooth Trainable Activation Functions
An activation function is a crucial component of a neural network that introduces non-linearity in the network. The state-of-the-art performance of a neural network depends also on the perfect choice of an activation function. We propose two novel non-monotonic smooth trainable activation functions, called ErfAct and Pserf. Experiments suggest that the proposed functions improve the network performance significantly compared to the widely used activations like ReLU, Swish, and Mish. Replacing ReLU by ErfAct and Pserf, we have 5.68% and 5.42% improvement for top-1 accuracy on Shufflenet V2 (2.0x) network in CIFAR100 dataset, 2.11% and 1.96% improvement for top-1 accuracy on Shufflenet V2 (2.0x) network in CIFAR10 dataset, 1.0%, and 1.0% improvement on mean average precision (mAP) on SSD300 model in Pascal VOC dataset.
- 12 January 2022
- Published by Springer Nature in Water Waves
- Vol. 4 (1) , 91-108
High-Frequency Instabilities of the Ostrovsky Equation
- 12 January 2022
- Published by Springer Nature in Water Waves
- Vol. 4 (1) , 91-108
High-Frequency Instabilities of the Ostrovsky Equation
- Open Access
- 16 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 120613-120623
TanhSoft—Dynamic Trainable Activation Functions for Faster Learning and Better Performance
Deep learning, at its core, contains functions that are the composition of a linear transformation with a non-linear function known as activation function. In the past few years, there is an increasing interest in the construction of novel activation functions resulting in better learning. In this work, we propose three novel activation functions with learnable parameters, namely TanhSoft-1, TanhSoft-2, and TanhSoft-3, which are shown to outperform several well-known activation functions. For instance, replacing ReLU with TanhSoft-1, TanhSoft-2, and Tanhsot-3 improves top-1 classification accuracy by 6.06%, 5.75%, and 5.38% respectively on VGG-16(with batch-normalization), by 3.02%, 3.25% and 2.93% respectively on PreActResNet-34 in CIFAR-100 dataset, by 1.76%, 1.93%, and 1.82% respectively on WideResNet 28-10 in Tiny ImageNet dataset. TanhSoft-1, TanhSoft-2, and Tanhsot-3 outperformed ReLU on mean average precision (mAP) by 0.7%, 0.8%, and 0.6% respectively in object detection problem on SSD 300 model in Pascal VOC dataset.
- 16 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 120613-120623
Open AccessTanhSoft—Dynamic Trainable Activation Functions for Faster Learning and Better Performance
Deep learning, at its core, contains functions that are the composition of a linear transformation with a non-linear function known as activation function. In the past few years, there is an increasing interest in the construction of novel activation functions resulting in better learning. In this work, we propose three novel activation functions with learnable parameters, namely TanhSoft-1, TanhSoft-2, and TanhSoft-3, which are shown to outperform several well-known activation functions. For instance, replacing ReLU with TanhSoft-1, TanhSoft-2, and Tanhsot-3 improves top-1 classification accuracy by 6.06%, 5.75%, and 5.38% respectively on VGG-16(with batch-normalization), by 3.02%, 3.25% and 2.93% respectively on PreActResNet-34 in CIFAR-100 dataset, by 1.76%, 1.93%, and 1.82% respectively on WideResNet 28-10 in Tiny ImageNet dataset. TanhSoft-1, TanhSoft-2, and Tanhsot-3 outperformed ReLU on mean average precision (mAP) by 0.7%, 0.8%, and 0.6% respectively in object detection problem on SSD 300 model in Pascal VOC dataset.