Research Publications
Browse peer-reviewed journal articles, conference papers, and scholarly publications. Use the filters below to refine results by type, author, year, or research area.
1.
Upadhyay, G.; Prasad, C. S.; Akash, V.; Singh, A. K.; Sai, Monish; Ghosal, S.; Kar, S.; Shantala, S. H.; Lalitkrushna, Thakur; Srividhya, S.; Balwantrao, S. A.
Institute of Electrical and Electronics Engineers Inc., 2024, ISBN: 979-835037289-2 (ISBN), (0).
Abstract | Links | Tags: OTHER
@proceedings{71,
title = {Enhancing CNN Inference Time and Reducing Latency on Edge and Resource-Constrained Systems through Quantization},
author = {G. Upadhyay and C. S. Prasad and V. Akash and A. K. Singh and Monish Sai and S. Ghosal and S. Kar and S. H. Shantala and Thakur Lalitkrushna and S. Srividhya and S. A. Balwantrao},
doi = {10.1109/NMITCON62075.2024.10699069},
isbn = {979-835037289-2 (ISBN)},
year = {2024},
date = {2024-01-01},
journal = {2nd IEEE International Conference on Networks, Multimedia and Information Technology, NMITCON 2024},
publisher = {Institute of Electrical and Electronics Engineers Inc.},
abstract = {Systems that use Deep Learning (DL) models extensively utilize cloud computing for inference tasks in various domains such as traffic monitoring, healthcare, and IoT. However, applications like autonomous vehicles, surveillance systems, and spacecraft are transitioning towards edge computing due to band-width limitations, transmission delays, and network connectivity issues. Edge computing mitigates these challenges by reducing latency through local data and model processing on the device. Implementing Deep Neural Networks (DNNs) on edge devices faces resource constraints, such as limited memory, computing power, etc. DNNs employ 32-bit floating-point precision for accuracy, leading to inflated model sizes. Quantization offers a solution by converting high-precision floating-point (FP) values to lower-precision or integer (INT) values, focusing on throughput and improving latency. This paper presents a comparative study of the accuracy and performance of 64-bit, 32-bit, and 16-bit floating-point instructions, along with 8-bit integer instructions, using Post Training Quantization (PTQ) and Quantization-Aware Training (QAT), on multiple Nets including CustomNets, which was inferenced on a GPU as well as a Xilinx Deep Processing Unit (DPU). The models were evaluated on a sample of the EuroSat Remote Sensing dataset. Quantizing models to FP16 and INT8 resulted in 2-3 x and 4 x faster inferencing, respectively, with a negligible decrease in accuracy of 1-4 %. FP64 exhibited a 2 3 x decrease in speed but a slight accuracy improvement (2 %). On the DPU, models showed minimal accuracy degradation of about 1 %. Overall, model size decreased by a constant 2 x and 4x from FP32 to FP16 and INT8, respectively, while increasing by 2 x for FP64. This reduction in size, with negligible loss in accuracy enables onboard storage along withfaster and accurate inferencing on resource constraint systems. © 2024 IEEE.},
note = {0},
keywords = {OTHER},
pubstate = {published},
tppubtype = {proceedings}
}
Systems that use Deep Learning (DL) models extensively utilize cloud computing for inference tasks in various domains such as traffic monitoring, healthcare, and IoT. However, applications like autonomous vehicles, surveillance systems, and spacecraft are transitioning towards edge computing due to band-width limitations, transmission delays, and network connectivity issues. Edge computing mitigates these challenges by reducing latency through local data and model processing on the device. Implementing Deep Neural Networks (DNNs) on edge devices faces resource constraints, such as limited memory, computing power, etc. DNNs employ 32-bit floating-point precision for accuracy, leading to inflated model sizes. Quantization offers a solution by converting high-precision floating-point (FP) values to lower-precision or integer (INT) values, focusing on throughput and improving latency. This paper presents a comparative study of the accuracy and performance of 64-bit, 32-bit, and 16-bit floating-point instructions, along with 8-bit integer instructions, using Post Training Quantization (PTQ) and Quantization-Aware Training (QAT), on multiple Nets including CustomNets, which was inferenced on a GPU as well as a Xilinx Deep Processing Unit (DPU). The models were evaluated on a sample of the EuroSat Remote Sensing dataset. Quantizing models to FP16 and INT8 resulted in 2-3 x and 4 x faster inferencing, respectively, with a negligible decrease in accuracy of 1-4 %. FP64 exhibited a 2 3 x decrease in speed but a slight accuracy improvement (2 %). On the DPU, models showed minimal accuracy degradation of about 1 %. Overall, model size decreased by a constant 2 x and 4x from FP32 to FP16 and INT8, respectively, while increasing by 2 x for FP64. This reduction in size, with negligible loss in accuracy enables onboard storage along withfaster and accurate inferencing on resource constraint systems. © 2024 IEEE.
2.
Balaji, N.; Pai, B. H.; Manjunath, K.; Venkatesh, B.; Bhavatarini, N.; Sreenidhi, B. K.
Cyberbullying in Online/E-Learning Platforms Based on Social Networks Proceedings
Springer Science and Business Media Deutschland GmbH, vol. 334, 2022, ISBN: 23673370 (ISSN); 978-981166368-0 (ISBN), (2).
Abstract | Links | Tags: OTHER
@proceedings{147,
title = {Cyberbullying in Online/E-Learning Platforms Based on Social Networks},
author = {N. Balaji and B. H. Pai and K. Manjunath and B. Venkatesh and N. Bhavatarini and B. K. Sreenidhi},
doi = {10.1007/978-981-16-6369-7_20},
isbn = {23673370 (ISSN); 978-981166368-0 (ISBN)},
year = {2022},
date = {2022-01-01},
journal = {Lecture Notes in Networks and Systems},
volume = {334},
pages = {227-240,},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {Cyberbullying is a method of harassment using electronic means and is very common among all, as the digital sphere has extended its technology. As school districts shut down in retort to COVID-19 and students are moving to E-learning, ever than before our children may be using most of the period in front of screens. Though the digital world is supporting us link and study more, it is also a situation where harmful behavior can be encountered. It is serious that we engross our children in dialogue to keep them protective and motivate them to be very vigilant with what they share online. To better safeguard your child, we wish you to study more about digital awareness, check your child’s screen time and online activities, comprehend and set settings related to privacy, and set rules with your child. Bullying is spiteful behavior that is aggressive, unwanted, and repeated. This can be verbal, social, physical, or online. They are acts of power and can directly cause guilt, sadness, shame, and anger. Bullying is not restricted to children; adults can also bully other adults or children. Cyberbullying is victimization that happens online and over digital devices. Cyberbullying examples comprise detestable or mean texts, social media posts envisioned to embarrassing or fake images, spread rumors, sexually explicit, or threatening direct messages. It is significant to take cyberbullying and bullying of any kind, seriously. Bullying can have a long-lasting effect on a child’s mental health, relationships, and confidence. It can distress their capability to focus on academics and extracurricular happenings. And, it can also cause a child to make to bullying others, as a technique to get control back and feel like they have power again. In social media where bigger communication opportunities are offered, they even enhance the defenselessness for many people in the form of threatening online with the messages, images, and so on. In recent studies, cyberbullying constitutes a growing problem among youngsters, teenagers, and school students. To avoid such situation, an intelligent system is required for the identification of these threats automatically. The chapter mainly focuses on the types of cyberbullying, followed by case studies in which we discuss about the automatic detection of cyberbullying in online learning/E-learning platform based on social network. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.},
note = {2},
keywords = {OTHER},
pubstate = {published},
tppubtype = {proceedings}
}
Cyberbullying is a method of harassment using electronic means and is very common among all, as the digital sphere has extended its technology. As school districts shut down in retort to COVID-19 and students are moving to E-learning, ever than before our children may be using most of the period in front of screens. Though the digital world is supporting us link and study more, it is also a situation where harmful behavior can be encountered. It is serious that we engross our children in dialogue to keep them protective and motivate them to be very vigilant with what they share online. To better safeguard your child, we wish you to study more about digital awareness, check your child’s screen time and online activities, comprehend and set settings related to privacy, and set rules with your child. Bullying is spiteful behavior that is aggressive, unwanted, and repeated. This can be verbal, social, physical, or online. They are acts of power and can directly cause guilt, sadness, shame, and anger. Bullying is not restricted to children; adults can also bully other adults or children. Cyberbullying is victimization that happens online and over digital devices. Cyberbullying examples comprise detestable or mean texts, social media posts envisioned to embarrassing or fake images, spread rumors, sexually explicit, or threatening direct messages. It is significant to take cyberbullying and bullying of any kind, seriously. Bullying can have a long-lasting effect on a child’s mental health, relationships, and confidence. It can distress their capability to focus on academics and extracurricular happenings. And, it can also cause a child to make to bullying others, as a technique to get control back and feel like they have power again. In social media where bigger communication opportunities are offered, they even enhance the defenselessness for many people in the form of threatening online with the messages, images, and so on. In recent studies, cyberbullying constitutes a growing problem among youngsters, teenagers, and school students. To avoid such situation, an intelligent system is required for the identification of these threats automatically. The chapter mainly focuses on the types of cyberbullying, followed by case studies in which we discuss about the automatic detection of cyberbullying in online learning/E-learning platform based on social network. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
3.
Kumar, M.; Kumar, D.; Akhtar, MdA. K.
A modified GA-based load balanced clustering algorithm for WSN: MGALBC Journal Article
In: International Journal of Embedded and Real-Time Communication Systems, vol. 12, pp. 44-63,, 2021, ISBN: 19473176 (ISSN), (12).
Abstract | Links | Tags: OTHER
@article{151,
title = {A modified GA-based load balanced clustering algorithm for WSN: MGALBC},
author = {M. Kumar and D. Kumar and MdA. K. Akhtar},
doi = {10.4018/IJERTCS.20210101.oa3},
isbn = {19473176 (ISSN)},
year = {2021},
date = {2021-01-01},
journal = {International Journal of Embedded and Real-Time Communication Systems},
volume = {12},
pages = {44-63,},
publisher = {IGI Global},
abstract = {The prevalent applications of WSN have fascinated a plethora of research efforts. Sensor nodes have serious limitations such as battery lifetime, memory constraints, and computational capabilities. Clustering is an important method for maximizing the network lifetime. In clustering, a network is divided into virtual groups, and CHs send their data to the BS either directly or using multi-hop routing. CHs are some special nodes having more energy than normal nodes. In fact, these special nodes are also battery operated and consequently power constrained; thus, they play a vital role in network lifetime. Cluster formation is very important and improper design may cause overload. This paper presents a modified GA-based load balanced clustering (MGALBC) algorithm for WSN. It is better than GA-based load balanced clustering (GALBC) algorithm because it balances the load by considering the residual energy. The result shows that the proposed method is better than GALBC in terms of energy consumption, number of active sensor nodes and network life. © 2021 IGI Global. All rights reserved.},
note = {12},
keywords = {OTHER},
pubstate = {published},
tppubtype = {article}
}
The prevalent applications of WSN have fascinated a plethora of research efforts. Sensor nodes have serious limitations such as battery lifetime, memory constraints, and computational capabilities. Clustering is an important method for maximizing the network lifetime. In clustering, a network is divided into virtual groups, and CHs send their data to the BS either directly or using multi-hop routing. CHs are some special nodes having more energy than normal nodes. In fact, these special nodes are also battery operated and consequently power constrained; thus, they play a vital role in network lifetime. Cluster formation is very important and improper design may cause overload. This paper presents a modified GA-based load balanced clustering (MGALBC) algorithm for WSN. It is better than GA-based load balanced clustering (GALBC) algorithm because it balances the load by considering the residual energy. The result shows that the proposed method is better than GALBC in terms of energy consumption, number of active sensor nodes and network life. © 2021 IGI Global. All rights reserved.
4.
Kantharaja, C. T.; Bharathi, V.
Awareness Of Vtu Consortium And Cloud-Based Online Resources In The Vtu Engineering College Libraries Journal Article
In: Library Philosophy and Practice, vol. 2020, 2020, ISBN: 15220222 (ISSN), (0).
@article{298,
title = {Awareness Of Vtu Consortium And Cloud-Based Online Resources In The Vtu Engineering College Libraries},
author = {C. T. Kantharaja and V. Bharathi},
isbn = {15220222 (ISSN)},
year = {2020},
date = {2020-01-01},
journal = {Library Philosophy and Practice},
volume = {2020},
publisher = {University of Idaho Library},
abstract = {Modern academic libraries should well be equipped with online resources along with print resources. Libraries are facing a transition period from print to electronic. The younger generation prefers to refer to the digital form of resources quickly. VTU Consortium is a blessing for readers, who are studying/working in VTU affiliated colleges. To meet the industrial requirements, one who updates his/her knowledge up to global and industrial standards. It can be achieved only by referring to current journals. VTU consortium has given that chance to library users to access world-class eJournals and eBooks on their campuses just by clicking the mouse. This study focuses on the usage of electronic resources subscribed through VTU consortium by students and teaching fraternity and how those resources being utilized by users. This paper gives the idea of promoting these resources at the college level. © 2020, Library Philosophy and Practice. All Rights Reserved.},
note = {0},
keywords = {OTHER},
pubstate = {published},
tppubtype = {article}
}
Modern academic libraries should well be equipped with online resources along with print resources. Libraries are facing a transition period from print to electronic. The younger generation prefers to refer to the digital form of resources quickly. VTU Consortium is a blessing for readers, who are studying/working in VTU affiliated colleges. To meet the industrial requirements, one who updates his/her knowledge up to global and industrial standards. It can be achieved only by referring to current journals. VTU consortium has given that chance to library users to access world-class eJournals and eBooks on their campuses just by clicking the mouse. This study focuses on the usage of electronic resources subscribed through VTU consortium by students and teaching fraternity and how those resources being utilized by users. This paper gives the idea of promoting these resources at the college level. © 2020, Library Philosophy and Practice. All Rights Reserved.
5.
Dean, R. V.; Ahmed, J.; Ahmed, S.; Indumathi, T. S.
New authentication architecture for GPRS networks Proceedings
2010, ISBN: 978-076953967-6 (ISBN), (1).
Abstract | Links | Tags: OTHER
@proceedings{320,
title = {New authentication architecture for GPRS networks},
author = {R. V. Dean and J. Ahmed and S. Ahmed and T. S. Indumathi},
doi = {10.1109/CISIS.2010.106},
isbn = {978-076953967-6 (ISBN)},
year = {2010},
date = {2010-01-01},
journal = {CISIS 2010 - The 4th International Conference on Complex, Intelligent and Software Intensive Systems},
pages = {649-654, 5447528+},
abstract = {GPRS offers the user can always on connection to Internet and Intranet. Some of the services may require high level of security. This can be financial transaction over the Internet, or exchange of confidential documents from a company's Intranet to an employee. It is important to have strong focus on the security, so companies and persons that demand high level of security can take advantage of the services GPRS offered. What normally happens is that the services win against the security. This is in most cases adequate security for what a normal subscriber requires. The possible attacks against GPRS targets the equipment of mobile users, the radio access network, the GPRS backbone network, and the interfaces that connect the latter to other GPRS networks or the public Internet. The results of these attacks might be the compromise of endusers security, the users over billing, the disclosure or alteration of critical information, the services unavailability, the network breakdown, etc. The analyzed attacks and their consequences increase the risks associated with the usage of GPRS, and, thus, influence its deployment that realizes the concept mobile Internet. In order to defeat certain attacks and enhance the level of security provided by GPRS, specific security measures need to be proposed. Hence there is a need of new authentication architecture which prevents malicious actions (attacks), which threaten the GPRS network, the GPRS mobile users, and the data that either reside at the network or are transferred through it. © 2010 IEEE.},
note = {1},
keywords = {OTHER},
pubstate = {published},
tppubtype = {proceedings}
}
GPRS offers the user can always on connection to Internet and Intranet. Some of the services may require high level of security. This can be financial transaction over the Internet, or exchange of confidential documents from a company's Intranet to an employee. It is important to have strong focus on the security, so companies and persons that demand high level of security can take advantage of the services GPRS offered. What normally happens is that the services win against the security. This is in most cases adequate security for what a normal subscriber requires. The possible attacks against GPRS targets the equipment of mobile users, the radio access network, the GPRS backbone network, and the interfaces that connect the latter to other GPRS networks or the public Internet. The results of these attacks might be the compromise of endusers security, the users over billing, the disclosure or alteration of critical information, the services unavailability, the network breakdown, etc. The analyzed attacks and their consequences increase the risks associated with the usage of GPRS, and, thus, influence its deployment that realizes the concept mobile Internet. In order to defeat certain attacks and enhance the level of security provided by GPRS, specific security measures need to be proposed. Hence there is a need of new authentication architecture which prevents malicious actions (attacks), which threaten the GPRS network, the GPRS mobile users, and the data that either reside at the network or are transferred through it. © 2010 IEEE.
6.
Indumathi, T. S.; Srinivas, T.; Rajaiah, K.; Rajappa, V.
Routing analysis and flexibility of ROADMs in all Optical networks Proceedings
2010, ISBN: 978-076954019-1 (ISBN), (2).
Abstract | Links | Tags: OTHER
@proceedings{337,
title = {Routing analysis and flexibility of ROADMs in all Optical networks},
author = {T. S. Indumathi and T. Srinivas and K. Rajaiah and V. Rajappa},
doi = {10.1109/WAINA.2010.92},
isbn = {978-076954019-1 (ISBN)},
year = {2010},
date = {2010-01-01},
journal = {24th IEEE International Conference on Advanced Information Networking and Applications Workshops, WAINA 2010},
pages = {75-78, 5480874+},
abstract = {Reconfigurable Optical Add / Drop Multiplexer is a subsystem of All Optical Networks. In this paper we analyze the routing performance of different ROADMs at dynamic traffic using Bernoulli, Poisson and Pascal distribution models. A theoretical routing power model is considered and through simulation results, the routing power and blocking probability have been evaluated to achieve better performance and flexibility. © 2010 IEEE.},
note = {2},
keywords = {OTHER},
pubstate = {published},
tppubtype = {proceedings}
}
Reconfigurable Optical Add / Drop Multiplexer is a subsystem of All Optical Networks. In this paper we analyze the routing performance of different ROADMs at dynamic traffic using Bernoulli, Poisson and Pascal distribution models. A theoretical routing power model is considered and through simulation results, the routing power and blocking probability have been evaluated to achieve better performance and flexibility. © 2010 IEEE.
7.
Biradar, A.; Thool, R. C.; Velur, R.
Voice transmission over local area network using bluetooth Proceedings
2009, ISBN: 978-142444547-9 (ISBN), (1).
Abstract | Links | Tags: OTHER
@proceedings{315,
title = {Voice transmission over local area network using bluetooth},
author = {A. Biradar and R. C. Thool and R. Velur},
doi = {10.1109/TENCON.2009.5395970},
isbn = {978-142444547-9 (ISBN)},
year = {2009},
date = {2009-01-01},
journal = {IEEE Region 10 Annual International Conference, Proceedings/TENCON},
pages = {5395970+},
abstract = {Bluetooth is a low cost, low power; short-range radio technology intended to transfer data between cell phones, PDAs and other portable devices. Bluetooth is also used as a medium for voice communication between cell phones. This paper aim is to develop an application that allows a user to make use of mobile phones enabled with Bluetooth and PC's connected over Local Area Network (LAN) in order to communicate with other user who is out of Bluetooth's range. ©2009 IEEE.},
note = {1},
keywords = {OTHER},
pubstate = {published},
tppubtype = {proceedings}
}
Bluetooth is a low cost, low power; short-range radio technology intended to transfer data between cell phones, PDAs and other portable devices. Bluetooth is also used as a medium for voice communication between cell phones. This paper aim is to develop an application that allows a user to make use of mobile phones enabled with Bluetooth and PC's connected over Local Area Network (LAN) in order to communicate with other user who is out of Bluetooth's range. ©2009 IEEE.
8.
Rajappa, V.; Prasad, K. V.
MHD flow and heat transfer in the flow of a power law fluid over a non-ISO thermal stretching sheet Proceedings
2008, ISBN: 978-076953267-7 (ISBN), (0).
Abstract | Links | Tags: OTHER
@proceedings{308,
title = {MHD flow and heat transfer in the flow of a power law fluid over a non-ISO thermal stretching sheet},
author = {V. Rajappa and K. V. Prasad},
doi = {10.1109/ICETET.2008.273},
isbn = {978-076953267-7 (ISBN)},
year = {2008},
date = {2008-01-01},
journal = {Proceedings - 1st International Conference on Emerging Trends in Engineering and Technology, ICETET 2008},
pages = {1077-1082, 4580063+},
abstract = {This article presents a numerical solution for the MHD flow of an electrically conducting non-Newtonian power law fluid over a semi-infinite non-isothermal stretching sheet with internal heat generation/ absorption. The flow is caused by linear stretching of a sheet from an impermeable wall. Thermal conductivity is assumed to vary linearly with temperature. The intricate coupled non-linear boundary value problem has been solved by a combination of Keller box method and shooting technique. An important observation of our study is that the velocity boundary layer thickness and thermal boundary layer thickness decreases with the increase of power law index in the presence / absence of variable thermal conductivity. © 2008 IEEE.},
note = {0},
keywords = {OTHER},
pubstate = {published},
tppubtype = {proceedings}
}
This article presents a numerical solution for the MHD flow of an electrically conducting non-Newtonian power law fluid over a semi-infinite non-isothermal stretching sheet with internal heat generation/ absorption. The flow is caused by linear stretching of a sheet from an impermeable wall. Thermal conductivity is assumed to vary linearly with temperature. The intricate coupled non-linear boundary value problem has been solved by a combination of Keller box method and shooting technique. An important observation of our study is that the velocity boundary layer thickness and thermal boundary layer thickness decreases with the increase of power law index in the presence / absence of variable thermal conductivity. © 2008 IEEE.
