Abhishek Kumar
@cuchd.in
Assistant Director
Chandigarh University
Dr. Abhishek Kumar is currently working as an Assistant director /Associate professor in Computer science & Engineering Department in Chandigarh University, Punjab, India .He is Doctorate in computer science from University of Madras and is doing Post-Doctoral Fellow in Ingenium Research Group Ingenium Research Group Lab, Universidad De Castilla-La Mancha, Ciudad Real, and Ciudad Real Spain. He has done M.Tech in Computer Sci. & Engineering and B.Tech in I.T. from, Rajasthan Technical University, Kota India. He has total Academic teaching experience of more than 11 years along with 2 years teaching assistantship. He is having more than 100 publications in reputed, peer reviewed National and International Journals, books & Conferences He has guided more than 30 M.Tech Projects at national and International level and guiding 6 PhD Scholar. His research area includes- Artificial intelligence, Renewable Energy Image processing, Computer Vision, Data Mining, Machine Learning. He has been Se
RESEARCH, TEACHING, or OTHER INTERESTS
Artificial Intelligence, Engineering, Health Information Management, Energy
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
P. Rihana Begum, Badrud Duza Mohammad, A. Saravana Kumar, and K.M. Muhasina
Wiley
S. Pradeep, Yogesh Kumar Sharma, Umesh Kumar Lilhore, Sarita Simaiya, Abhishek Kumar, Sachin Ahuja, Martin Margala, Prasun Chakrabarti, and Tulika Chakrabarti
Springer Science and Business Media LLC
AbstractSoftware-defined networking (SDN) has significantly transformed the field of network management through the consolidation of control and provision of enhanced adaptability. However, this paradigm shift has concurrently presented novel security concerns. The preservation of service path integrity holds significant importance within SDN environments due to the potential for malevolent entities to exploit network flows, resulting in a range of security breaches. This research paper introduces a model called "EnsureS", which aims to enhance the security of SDN by proposing an efficient and secure service path validation approach. The proposed approach utilizes a Lightweight Service Path Validation using Batch Hashing and Tag Verification, focusing on improving service path validation's efficiency and security in SDN environments. The proposed EnsureS system utilizes two primary techniques in order to validate service pathways efficiently. Firstly, the method utilizes batch hashing in order to minimize computational overhead. The proposed EnsureS algorithm enhances performance by aggregating packets through batches rather than independently; the hashing process takes place on each one in the service pathway. Additionally, the implementation of tag verification enables network devices to efficiently verify the authenticity of packets by leveraging pre-established trust relationships. EnsureS provides a streamlined and effective approach for validating service paths in SDN environments by integrating these methodologies. In order to assess the efficacy of the Proposed EnsureS, a comprehensive series of investigations were conducted within a simulated SDN circumstance. The efficacy of Proposed EnsureS was then compared to that of established methods. The findings of our study indicate that the proposed EnsureS solution effectively minimizes computational overhead without compromising on the established security standards. The implementation successfully reduces the impact of different types of attacks, such as route alteration and packet spoofing, increasing SDN networks' general integrity.
Rahul Kumar, Shweta Singh, Shweta Chauhan, Abhineet Anand, and Abhishek Kumar
Elsevier BV
Osama Bassam J. Rabie, Shitharth Selvarajan, Daniyal Alghazzawi, Abhishek Kumar, Syed Hasan, and Muhammad Zubair Asghar
Institution of Engineering and Technology (IET)
AbstractDetection of cyber‐threats in the smart grid Supervisory Control and Data Acquisition (SCADA) is still remains one of the complex and essential processes need to be highly concentrated in present times. Typically, the SCADA is more prone to the security issues due to their environmental problems and vulnerabilities. Therefore, the proposed work intends to design a new detection approach by integrating the optimization and classification models for smart grid SCADA security. In this framework, the min‐max normalization is performed at first for noise removal and attributes arrangement. Here, the correlation estimation mechanism is mainly deployed to reduce the dimensionality of features by choosing the relevant features used for attack prediction. Moreover, the optimal features are selected by using the optimal solution provided by the Holistic Harris Hawks Optimization (H3O). Finally, the Perceptron Stochastic Neural Network (PSNN) is utilized to categorize the normal and attacking data flow in the network with minimal processing time and complexity. By using the combination of proposed H3O‐PSNN technique, the detection accuracy is improved up to 99% for all datasets used in this study, and also other measures such as precision to 99.2%, recall to 99%, f1‐score to 99.2% increased, when compared to the standard techniques.
Abhishek Kumar, Pramod Sing Rathore, Ashutosh Kumar Dubey, Rashmi Agrawal, and Kanta Prasad Sharma
Springer Science and Business Media LLC
Abhishek Kumar, Pramod Sing Rathore, Ashutosh Kumar Dubey, Rashmi Agrawal, and Kanta Prasad Sharma
Springer Science and Business Media LLC
Abhishek Kumar, Swarn Avinash Kumar, Vishal Dutt, S. Shitharth, and Esha Tripathi
Wiley
Ilyas Benkhaddra, Abhishek Kumar, Mohamed Ali Setitra, and Lei Hang
Springer Science and Business Media LLC
Arun Lal Srivastav, Markandeya, Naveen Patel, Mayank Pandey, Ashutosh Kumar Pandey, Ashutosh Kumar Dubey, Abhishek Kumar, Abhishek Kumar Bhardwaj, and Vinod Kumar Chaudhary
Springer Science and Business Media LLC
Benkhaddra Ilyas, Abhishek Kumar, Mohamed Ali Setitra, ZineEl Abidine Bensalem, and Hang Lei
Wiley
The attack named Distributed Denial of Service (DDoS) that takes place in the large blockchain network requires an efficient and robust attack detection and prevention mechanism for authenticated access. Blockchain is a distributed network in which the attacker tries to hack the network by utilizing all the resources with the application of enormous requests. Several methods like Rival Technique, filter modular approach and so on, were developed to detect and prevent the DDoS attack in the blockchain; still, detection accuracy is a challenging task. Hence, this research introduces an efficient technique using optimization‐based deep learning by considering the blockchain network and smart contract for the detection and prevention of DDoS attacks. Based on the user request, the traffic is analyzed, and the verification using the smart contract is made to find the authenticated user. After the verification, the response is provided for the authenticated user, and the suspicious traffic is utilized for the detection of DDoS attacks using the Poaching Raptor Optimization‐based deep neural network (Poaching Raptor‐based DNN), in which the classifier is tuned using the proposed optimization algorithm to reduce the training loss. The proposed algorithm is designed by hybridizing the habitual practice of the raptor by considering the concurring behavior, hunting style along with poaching behavior of the Lobo to enhance the detection accuracy. After the attack detection, the nonattacker is responded, and the attacker is prevented by entering the IP/MAC address in the logfile. The performance of the proposed method is evaluated in terms of recall, precision, FPR, and accuracy and obtained the values of 96.3%, 98.22%, 3.33%, and 95.12%, respectively.
Venkata Raghuveer Burugadda, Prashant S. Pawar, Abhishek Kumar, and Neha Bhati
IEEE
Heart failure is a frequent cause of hospitalization and readmission because of the severity of the disease. Researchers explored using Machine Learning (ML) algorithms to forecast whether heart failure patients must be readmitted to the hospital. This study examines ML algorithms that use data from electronic health records to forecast hospital readmissions for patients with heart failure. We will assess the accuracy, precision, recall, and F1-score of logistic regression, decision trees, random forests, Support Vector Machines (SVM), and artificial neural networks. The study's results will show how well ML algorithms predict heart failure patients' hospital readmission risk, which could lead to personalized therapies that improve patient outcomes and save healthcare costs. Comparing studies in this field shows gaps in model interpretability, generalizability, and socioeconomic determinants of health in prediction models.
Venkata Raghuveer Burugadda, Priyanka Makrand Mane, Abhishek Kumar, and Neha Bhati
IEEE
Early sepsis detection improves patient outcomes and care. This research provides a Machine Learning (ML) system for hospitalized sepsis detection. Gradient boosting, an ensemble learning method, analyses patient data to detect sepsis early. A comprehensive electronic health record database, MIMIC-III, was used to design and test the algorithm. The algorithm's sepsis detection accuracy, precision, recall, F1 score, and ROC AUC were measured. The proposed approach was more accurate than traditional models. It accurately predicted sepsis patients and aid treatment. Real-time clinical decision-making is possible with the algorithm's fast prediction and training. It could revolutionize sepsis management by giving doctors a dependable early detection and intervention tool. The algorithm must be tested in various healthcare environments and patient demographics. To implement this technology widely, privacy and ethics must be addressed. The approach may improve patient outcomes and lower healthcare costs by detecting sepsis early.
Oskar Krishna Shrestha, Sakar Khatiwada, Adarsha Ghimire, Bibhuti Rajbhandari, and A Vijay Kumar
IEEE
This research paper suggests a machine learning-based system designed to interpret sign languages in real-time from a video feed to help deaf and hard of hearing individuals overcome communication barriers. Sign languages employ hand gestures, facial expressions, and body language to convey ideas and meaning visually. Unfortunately, individuals with hearing impairments face communication difficulties, limited access to information and services, and social isolation. The proposed system captures the subject's gestures using a camera, which are then processed using a TensorFlow object detection API that predicts based on a pre-trained machine learning model. The system's methodology incorporates supervised learning, Single Shot MultiBox Detector (SSD) object detection algorithm, and the luminosity method for converting colour images to grayscale. For dataset creation, hand gesture image samples were captured and labelled using open-source graphical annotation software LabelImg, and the model was trained. The system proposed achieved an average accuracy of 95% and has the potential to improve communication and reduce marginalization for deaf and hard of hearing individuals.
Rajasrikar Punugoti, Vishal Dutt, Abhishek Kumar, and Neha Bhati
IEEE
Cardiovascular Disease (CVD) affects deaths and hospitalisations. Clinical data analytics struggles to predict heart disease survival. This report compares machine learning-based cardiovascular disease prediction studies. The authors use a Kaggle dataset of 70,000 records and 16 features to show a SMOTE model with hyperparameter-optimized classifiers. Random Forest outperforms KNN with 13 elements in cardiovascular disease prediction. Naive Bayes outperforms SVM on complete feature sets. The proposed model achieves 86% accuracy, and the optimised SMOTE technique outperforms the traditional SMOTE technique in all metrics. This study analyses the strengths and weaknesses of existing models for making cardiovascular disease predictions with machine learning and suggests a promising new method.
Shagun Preet Kour, Abhishek Kumar, and Sachin Ahuja
IEEE
Early identification of diabetes is vital since it's an incurable condition with no complete cure. We used data mining and machine learning strategies in our investigation to anticipate diabetes. 768 individuals and their relevant attributes are the focus of the hour. Few machine learning methods have been applied to the dataset for the goal of forecasting the occurrence of diabetes. The implemented algorithms' consistency and harshness have been investigated using methods that focused on correlating accuracy and F-1 rankings. Comparison between algorithms is done to increase the accuracy in comparison. We predicted Extra Tree Classifier gives 80% accuracy compared to support vector Machine. The goal of this study is to create a feasible strategy that will aid medical personnel with the diagnosis of diabetic complications at a tender point.
Uday Shankar Sekhar, Narayan Vyas, Vishal Dutt, and Abhishek Kumar
IEEE
This research aimed to evaluate numerous deep-learning models for Alzheimer's disease detection using several different neuroimaging techniques. Ten recent studies were selected for comparison based on their methodology, conclusions, and limitations. The Generative Adversarial Network (GAN) algorithm is applied fictitiously to the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, and results are provided. A comparison was made between the results of the GAN algorithm and the selected studies. Evaluation metrics were presented in a table and a graph. The study concludes that ensemble models and multi-modal approaches improve Alzheimer's Disease detection performance. However, there is a need for further work to be done to address issues, including limited sample sizes and a lack of external validation.
Rishabh Dhiman, Anupam Baliyan, and Abhishek Kumar
IEEE
Face spoof detection is a scheme for detecting the spoofed faces. Various techniques have been designed for the face spoof detection, based on features, machine learning (ML) and deep learning (DL) in the past few years, and has drawn great interest in a variety of domains that are most accessible and leads to maintain abundant information in day-to-day life. Face spoof detection techniques which are based on Artificial Intelligence (AI) contain several stages such as pre-process the data, extract the features and classify the faces. The earlier method makes the deployment of K-Nearest Neighbor with Grey Level Co-occurrence Matrix to detect the face spoofing, but at lower accuracy. Moreover, $\\text{DWT}-\\text{LBP}-\\text{DCT}+\\text{SVM}$ algorithm is also implemented to detect the face-spoofing. However, the accuracy of these algorithms is found lower. This work projects a hybrid technique in which RF, K-Nearest Neighbor (KNN) and Support Vector Machine algorithms are put together. The simulation results exhibited the supremacy of the projected technique as compared to other methods concerning accuracy.
Priya Batta, Sachin Ahuja, and Abhishek Kumar
IEEE
Many industries have serious concerns about the secure transmission of sensitive data, and the use of blockchain technology, steganography, and optimization algorithms has emerged as a potent remedy. Steganography and optimization algorithms can hide data and increase the effectiveness of the transmission process, while blockchain technology offers a transparent and secure platform for data storage and transfer. The employment of these technologies in tandem to offer a reliable and secure solution for the transfer of sensitive data is provided in this paper. We also go over this strategy's possible advantages and difficulties and offer suggestions for successfully putting it into practice. Overall, a promising method for the secure transfer of confidential data in the current digital era is provided by the combination of blockchain, steganography, and optimization algorithms. This tactic ensures the development of a reliable and efficient blockchain platform for the management of sensitive data while upholding predetermined standards.
Manohar Kapse, N. Elangovan, Abhishek Kumar, and Joseph Durai Selvam
Springer Nature Singapore
Abhishek Kumar, Bashant Kumar Sah, Tushar Mehrotra, and Gaurav Kumar Rajput
IEEE
The double-spending problem in blockchain technology is a significant challenge that threatens the integrity and trustworthiness of decentralized systems. This problem occurs when a user attempts to spend the same cryptocurrency unit twice, leading to a situation where the blockchain network must decide which transaction to accept and which to reject. One of the most urgent problems with blockchain technology is the issue of double spending, as it undermines the fundamental principles of trust and transparency that underlie decentralized systems. Various factors can contribute to the double-spending problem, including network latency, malicious actors, and the consensus mechanism used to validate transactions. This study investigates the many approaches put out to solve the double-spending issue in blockchain technology. The proof-of-work consensus mechanism, which necessitates network users to carry out difficult calculations in order to validate transactions, is one of the most popular alternatives. The proof-of-stake consensus technique is an additional remedy, which relies on participants staking their cryptocurrency units to validate transactions. While both mechanisms have their advantages and disadvantages, they are not foolproof and can be vulnerable to attacks. Emerging technologies, like multi-party computation and zero-knowledge proofs, are being investigated in addition to current solutions to the double-spending issue. Overall, this paper highlights the critical nature of the double-spending problem in blockchain technology and evaluates the existing and emerging solutions to the issue.
RECENT SCHOLAR PUBLICATIONS
MOST CITED SCHOLAR PUBLICATIONS
RESEARCH OUTPUTS (PATENTS, SOFTWARE, PUBLICATIONS, PRODUCTS)
Method Of Data Transmission In A Cluster Network
INDIAN PATENT OFFICE
System And Method For Cluster Head Selection And Cluster Formation For Improving Radio Frequency Identification
Network
INDIAN PATENT OFFICE
202111022269
Iot Enabled Wall Climbing Robot For Security
IP AUSTRALIA /GRANTED
2021101471
An Artificial Intelligence And IoT Based Method For Prevention Of Security Attack On Cloud Medical Data
IP AUSTRALIA/ GRANTED
2021102115
Iot Based Generic Framework For Computer Security Using Artificial Immune System
IP AUSTRALIA /GRANTED
2021102104
Podium with display facility, box and glass holder
INDIAN PATENT OFFICE/GRANTED
346057-001
Hexa Tube LED Bulb
INDIAN PATENT OFFICE/GRANTED
356883001
SMART SHOPPING CART
INDIAN PATENT OFFICE
202111061690
202111018897