Jaypee Institute of Information Technology

About: Jaypee Institute of Information Technology is a education organization based out in Noida, Uttar Pradesh, India. It is known for research contribution in the topics: Computer science & Cluster analysis. The organization has 2136 authors who have published 3435 publications receiving 31458 citations. The organization is also known as: JIIT Noida.

A Hybrid Model for Music Genre Classification Using LSTM and SVM

Abstract: With today's cutting edge technology and intractable access to voluminous data files via the internet, it is important to meet the computational needs of every user. Machine learning is one such growing branch of artificial intelligence that has made such demands of the users viable. Machine learning models are paving the way for classification techniques such as in music genre classification, and have shown to be efficient in predicting classes to a great extent. To exploit the time dependent nature of the dataset Long Short-Term Memory (LSTM) Neural Network is used for music genre classification and combined with Support Vector Machine (SVM) classifier to enhance its performance. The hybrid model of these two classifiers resulted into an increase in the accuracy of prediction of the individual models. This hybrid model is imposed on GTZAN music dataset and is compared with the results of standalone models of LSTM and SVM. The proposed model exceeded the independent accuracies of the LSTM and SVM classifiers with an accuracy of 89%, reaffirming the efficient utilization of each classifier.

A new recommendation system using map-reduce-based tournament empowered Whale optimization algorithm

Abstract: In the era of Web 2.0, the data are growing immensely and is assisting E-commerce websites for better decision-making. Collaborative filtering, one of the prominent recommendation approaches, performs recommendation by finding similarity. However, this approach fails in managing large-scale datasets. To mitigate the same, an efficient map-reduce-based clustering recommendation system is presented. The proposed method uses a novel variant of the whale optimization algorithm, tournament selection empowered whale optimization algorithm, to attain the optimal clusters. The clustering efficiency of the proposed method is measured on four large-scale datasets in terms of F-measure and computation time. The experimental results are compared with state-of-the-art map-reduce-based clustering methods, namely map-reduce-based K-means, map-reduce-based bat algorithm, map-reduce-based Kmeans particle swarm optimization, map-reduce-based artificial bee colony, and map-reduce-based whale optimization algorithm. Furthermore, the proposed method is tested as a recommendation system on the publicly available movie-lens dataset. The performance validation is measured in terms of mean absolute error, precision and recall, over a different number of clusters. The experimental results assert that the proposed method is a permissive approach for the recommendation over large-scale datasets.

ACE2 Homo-dimerization, Human Genomic variants and Interaction of Host Proteins Explain High Population Specific Differences in Outcomes of COVID19

Abstract: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a positive single stranded RNA virus that causes a highly contagious Corona Virus Disease (COVID19). Entry of SARS-CoV-2 in human cells depends on binding of the viral spike (S) proteins to cellular receptor Angiotensin-converting enzyme 2 (ACE2) and on S protein priming by host cell serine protease TMPRSS2. Recently COVID19 has been declared pandemic by World Health Organization yet high differences in disease outcomes across countries have been seen. We provide evidences based on analyses of existing public datasets and by using various in-silico approaches to explain some of these as factors that may explain population level differences. One of the key factors might be entry of virus in host cells due to differential interaction of viral proteins with host cell proteins due to different genetic backgrounds. Based on our findings, we conclude that higher expression of ACE2 facilitated by natural variations, acting as Expression quantitative trait loci (eQTLs) and with different frequencies in different populations, results in ACE2 homo-dimerization which is disadvantageous for TMPRSS2 mediated cleavage of ACE2 and becomes more difficult in presence of broad neutral amino acid transporter, B0AT1 (coded by SLC6A19), that usually does not express in Lungs. We also propose that the monomeric ACE2 has higher preferential binding with SARS-CoV-2 S-Protein vis-a-vis its dimerized counterpart. Further, eQTLs in TMPRSS2 and natural structural variations in the gene may also result in differential outcomes towards priming of viral S-protein, a critical step for entry of Virus in host cells. In addition, we suggest some other potential key host genes like ADAM17, RPS6, HNRNPA1, SUMO1, NACA, BTF3 and some other proteases as Cathepsins, that might have a critical role. Understanding these population specific differences may help in developing appropriate management strategies.

Convolutional Neural Network Based American Sign Language Static Hand Gesture Recognition

Abstract: Communicating through hand gestures with each other is simply called the language of signs. It is an acceptable language for communication among deaf and dumb people in this society. The society of...

Low Power Techniques for Digital System Design

Abstract: The proliferation of reconfigurable hardware like (FPGAs) put a challenge in front of designers to implement fast and low powered digital designs. Main drawbacks of FPGAs are the complex circuitry which makes them less efficient as compared to ASIC (Application Specific Integrated Circuits). Although appropriate to scaling in CMOS technology reduce the power required for performing the known job, it increase clout indulgence for each part of region. At similar instant request of low power application is swelling due to increase of smart devices and increasing energy costs. Since power consumption is an extremely significant issue in digital classification of designs, so the authors have presented and analyzed some power reduction techniques that can be targeted at different levels of design hierarchy for different target platform. The authors would also discuss concept of ACPI module designed for newer operating systems, which provides basic power management facilities to save system power.