Suraj Kumar Pandey
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Suraj Kumar Pandey has a total of 2 co-authors. Here are the top co-authors by number of shared publications:
- Shivashankar Bhaskaran Nair
Indian Institute of Technology Guwahati
6 shared publications - Karthik Mohan
SASTRA Deemed University
1 shared publication
Latest Publications
- Open Access
- 5 December 2025
- Published by Association for Computing Machinery (ACM) in ACM Journal on Emerging Technologies in Computing Systems
- Vol. 22 (1) , 1-22
ChaoticImmuneNet: A Chaos-driven Immunity Inspired Neural Network Paradigm for Embodied Intelligence in Resource-Constrained Devices
Edge devices (ED) play a crucial role in enabling intelligent decision-making capabilities for Smart Cyber-Physical Systems (CPS). Nevertheless, it is a challenging objective, given the resource constraint of edge devices. Pre-trained Machine learning models fail to perform well against real-world data, provided via EDs. Conventional Embodied AI strategies outsource the training load to adapt to real-world data. However, the dependence on outsourcing attracts issues related to latency, privacy, scalability, and so on that are detrimental to a CPS. As a solution, this c proposes ChaoticImmuneNet , a lightweight Embodied AI method that allows onboard learning on resource-constrained EDs, with limited and noisy data. The article introduces a novel strategy, merging techniques from Artificial Immune Systems and Chaos Theory with a Siamese Neural Network for realizing onboard learning. ChaoticImmuneNet differs from Chaotic Neural Networks as the former does not require specialized models and hardware to leverage chaotic dynamics. Experimental studies carried out using three diverse image datasets demonstrate the efficacy of the proposed method when compared with existing onboard learning techniques, in terms of accuracy, time and storage requirements. A real-world deployment of the ChaoticImmuneNet on a real mobile robot, operating within a warehouse prototype testify to its pragmatic utility.
- 5 December 2025
- Published by Association for Computing Machinery (ACM) in ACM Journal on Emerging Technologies in Computing Systems
- Vol. 22 (1) , 1-22
Open AccessChaoticImmuneNet: A Chaos-driven Immunity Inspired Neural Network Paradigm for Embodied Intelligence in Resource-Constrained Devices
Edge devices (ED) play a crucial role in enabling intelligent decision-making capabilities for Smart Cyber-Physical Systems (CPS). Nevertheless, it is a challenging objective, given the resource constraint of edge devices. Pre-trained Machine learning models fail to perform well against real-world data, provided via EDs. Conventional Embodied AI strategies outsource the training load to adapt to real-world data. However, the dependence on outsourcing attracts issues related to latency, privacy, scalability, and so on that are detrimental to a CPS. As a solution, this c proposes ChaoticImmuneNet , a lightweight Embodied AI method that allows onboard learning on resource-constrained EDs, with limited and noisy data. The article introduces a novel strategy, merging techniques from Artificial Immune Systems and Chaos Theory with a Siamese Neural Network for realizing onboard learning. ChaoticImmuneNet differs from Chaotic Neural Networks as the former does not require specialized models and hardware to leverage chaotic dynamics. Experimental studies carried out using three diverse image datasets demonstrate the efficacy of the proposed method when compared with existing onboard learning techniques, in terms of accuracy, time and storage requirements. A real-world deployment of the ChaoticImmuneNet on a real mobile robot, operating within a warehouse prototype testify to its pragmatic utility.
- 25 April 2025
- Published by Elsevier in Internet of Things
- Vol. 32
: A Virtual-to-Real knowledge transfer and adaptation method for edge devices- 25 April 2025
- Published by Elsevier in Internet of Things
- Vol. 32
: A Virtual-to-Real knowledge transfer and adaptation method for edge devices- 2 January 2025
- Published by Taylor & Francis in IETE Technical Review
- Vol. 42 (1) , 148-161
RoboDA: A Dataset for Domain Adaptation in Robot Vision
- 2 January 2025
- Published by Taylor & Francis in IETE Technical Review
- Vol. 42 (1) , 148-161
RoboDA: A Dataset for Domain Adaptation in Robot Vision
- 4 December 2024
- Published by Springer Nature
- p. 226-242
A Deep-Learning Based Real-Time License Plate Recognition System for Resource-Constrained Scenarios
- 4 December 2024
- Published by Springer Nature
- p. 226-242
A Deep-Learning Based Real-Time License Plate Recognition System for Resource-Constrained Scenarios
- Open Access
- 14 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 299-302
Onboard Class Incremental Learning for Resource-Constrained scenarios using Genetic Algorithm and TinyML
Deploying Machine Learning (ML) models in real-world settings over resource-constrained edge devices has always been a challenging task. While TinyML tackles this issue to an extent, by mostly using pre-trained Deep Learning (DL) models, the static nature of such models renders them ineffective for non-stationary data. A model having a low memory footprint that could allow onboard Class Incremental Learning (CIL) so as to accommodate data from new classes and also avoid catastrophic forgetting, is thus, the need of the day. The work described in this paper endeavours to meet this need by providing a method that utilises TinyML to accommodate a DL model onboard a resource-constrained device. To enable onboard CIL over the DL model, the method leverages Latent Replays (as exemplars) and a Genetic Algorithm (GA) to create a multi-fitness landscape that facilitates the inclusion of new class data, suppresses catastrophic forgetting and keeps a check on the quality of exemplars. Experiments and comparisons conducted for a gesture recognition task using time-series data with the proposed method deployed on a microcontroller, show the effectiveness of augmenting TinyML DL models with the GA for onboard CIL.
- 14 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 299-302
Open AccessOnboard Class Incremental Learning for Resource-Constrained scenarios using Genetic Algorithm and TinyML
Deploying Machine Learning (ML) models in real-world settings over resource-constrained edge devices has always been a challenging task. While TinyML tackles this issue to an extent, by mostly using pre-trained Deep Learning (DL) models, the static nature of such models renders them ineffective for non-stationary data. A model having a low memory footprint that could allow onboard Class Incremental Learning (CIL) so as to accommodate data from new classes and also avoid catastrophic forgetting, is thus, the need of the day. The work described in this paper endeavours to meet this need by providing a method that utilises TinyML to accommodate a DL model onboard a resource-constrained device. To enable onboard CIL over the DL model, the method leverages Latent Replays (as exemplars) and a Genetic Algorithm (GA) to create a multi-fitness landscape that facilitates the inclusion of new class data, suppresses catastrophic forgetting and keeps a check on the quality of exemplars. Experiments and comparisons conducted for a gesture recognition task using time-series data with the proposed method deployed on a microcontroller, show the effectiveness of augmenting TinyML DL models with the GA for onboard CIL.
- Open Access
- 24 March 2024
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 38 (21) , 23608-23609
Neuroevolution of a Multi-Generator GAN (Student Abstract)
Evolutionary Algorithms (EA) have been leveraged to tackle the challenges faced while using GANs such as mode collapse, vanishing gradient, latent space search, etc. However, the existing techniques of using EA with GANs operate backpropagation and EA in isolation from each other, leaving ample room for further exploration. This paper creates a collaborative bridge between EA and GANs by exploring a neuroevolution method for utilising both EA and backpropagation-based optimisation, simultaneously, for a multi-generator GAN architecture. Experiments conducted using a standard dataset with variants of the proposed method highlight the towering impact of each of the components involved in the proposed method.
- 24 March 2024
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 38 (21) , 23608-23609
Open AccessNeuroevolution of a Multi-Generator GAN (Student Abstract)
Evolutionary Algorithms (EA) have been leveraged to tackle the challenges faced while using GANs such as mode collapse, vanishing gradient, latent space search, etc. However, the existing techniques of using EA with GANs operate backpropagation and EA in isolation from each other, leaving ample room for further exploration. This paper creates a collaborative bridge between EA and GANs by exploring a neuroevolution method for utilising both EA and backpropagation-based optimisation, simultaneously, for a multi-generator GAN architecture. Experiments conducted using a standard dataset with variants of the proposed method highlight the towering impact of each of the components involved in the proposed method.
- Open Access
- 15 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 39-40
Enhancing Siamese Neural Networks for Multi-class Classification: An Immuno-inspired approach
Siamese Neural Networks (SNN) are known to perform well in resource-constrained scenarios where computation and data availability are limited. They utilise the similarity space of a given dataset to extract distinguishing features between dissimilar data samples. Such features have also been utilised for classification tasks. Though several works on enhancing the accuracies and inference times using such similarity spaces have been reported, there is still scope for investigations that can yield more efficient strategies. The Biological Immune System (BIS) is known for employing such a transformation to recognise and contain antigenic attacks. Concepts from a BIS can thus, aid in boosting the classification performance of SNNs. This paper summarizes such an attempt made in our work "Immuno-Inspired Augmentation of Siamese Neural Network for Multi-class Classification" [8] presented at IVCNZ 2022, first published in Lecture Notes in Computer Science, 2023, vol 13836, pages 486--500 by Springer. A novel SNN-based multi-class classification method augmented with an immuno-inspired approach that allows an SNN to plug class-specific characteristics into its architecture is presented herein. The empirical analyses and results conducted on three benchmark datasets, clearly indicate that this method delivers higher accuracies and lower inference times when compared to recent SNN-based multi-class classification techniques.
- 15 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 39-40
Open AccessEnhancing Siamese Neural Networks for Multi-class Classification: An Immuno-inspired approach
Siamese Neural Networks (SNN) are known to perform well in resource-constrained scenarios where computation and data availability are limited. They utilise the similarity space of a given dataset to extract distinguishing features between dissimilar data samples. Such features have also been utilised for classification tasks. Though several works on enhancing the accuracies and inference times using such similarity spaces have been reported, there is still scope for investigations that can yield more efficient strategies. The Biological Immune System (BIS) is known for employing such a transformation to recognise and contain antigenic attacks. Concepts from a BIS can thus, aid in boosting the classification performance of SNNs. This paper summarizes such an attempt made in our work "Immuno-Inspired Augmentation of Siamese Neural Network for Multi-class Classification" [8] presented at IVCNZ 2022, first published in Lecture Notes in Computer Science, 2023, vol 13836, pages 486--500 by Springer. A novel SNN-based multi-class classification method augmented with an immuno-inspired approach that allows an SNN to plug class-specific characteristics into its architecture is presented herein. The empirical analyses and results conducted on three benchmark datasets, clearly indicate that this method delivers higher accuracies and lower inference times when compared to recent SNN-based multi-class classification techniques.
- 4 February 2023
- Published by Springer Nature
- p. 486-500
Immuno-Inspired Augmentation of Siamese Neural Network for Multi-class Classification
- 4 February 2023
- Published by Springer Nature
- p. 486-500
Immuno-Inspired Augmentation of Siamese Neural Network for Multi-class Classification