Kashif Bilal

Dr. Kashif Bilal


Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 October 2018 in IEEE Transactions on Parallel and Distributed Systems
Verified
Journal Article
Closed access
JCJianguo Chen
Jianguo Chen
KLKenli Li
Kenli Li
KBKashif Bilal
Kashif Bilal
XZXu Zhou
Xu Zhou
Keqin LiKeqin Li
Keqin Li
PYPhilip S. Yu
Philip S. Yu
Benefitting from large-scale training datasets and the complex training network, Convolutional Neural Networks (CNNs) are widely applied in various fields with high accuracy. However, the training process of CNNs is very time-consuming, where large amounts of training samples and iterative operations are required to obtain high-quality weight parameters. In this paper, we focus on the time-consuming training process of large-scale CNNs and propose a Bi-layered Parallel Training (BPT-CNN) architecture in distributed computing environments. BPT-CNN consists of two main components: (a) an outer-layer parallel training for multiple CNN subnetworks on separate data subsets, and (b) an inner-layer parallel training for each subnetwork. In the outer-layer parallelism, we address critical issues of distributed and parallel computing, including data communication, synchronization, and workload balance. A heterogeneous-aware Incremental Data Partitioning and Allocation (IDPA) strategy is proposed, where large-scale training datasets are partitioned and allocated to the computing nodes in batches according to their computing power. To minimize the synchronization waiting during the global weight update process, an Asynchronous Global Weight Update (AGWU) strategy is proposed. In the inner-layer parallelism, we further accelerate the training process for each CNN subnetwork on each computer, where computation steps of convolutional layer and the local weight training are parallelized based on task-parallelism. We introduce task decomposition and scheduling strategies with the objectives of thread-level load balancing and minimum waiting time for critical paths. Extensive experimental results indicate that the proposed BPT-CNN effectively improves the training performance of CNNs while maintaining the accuracy.

17 October 2017 in IEEE Transactions on Big Data
Verified
Journal Article
Closed access
HLHuazhong Liu
Huazhong Liu
BLBaoshun Liu
Baoshun Liu
Laurence YangLaurence Yang
Laurence Yang
MLMan Lin
Man Lin
YDYuhui Deng
Yuhui Deng
SKSamee U. Khan
Samee U. Khan
Big data has received considerable attention in recent years because of massive data volumes in various fields. Considering various “V” features, big data tasks are usually highly complex and computational intensive. These tasks are generally performed in parallel in data centers resulting in massive energy consumption and Green House Gases emissions. Therefore, efficient resource allocation considering the synergy of the performance and energy efficiency is one of the crucial challenges today. In this paper, we aim to achieve maximum energy efficiency by combining thermal-aware and Dynamic Voltage and Frequency Scaling techniques. This paper proposes: (a) a thermal-aware and power-aware hybrid energy consumption model synchronously considering the computing, cooling, and migration energy consumption; (b) a tensor-based task allocation and frequency assignment representation model for representing the relationship among different tasks, nodes, time slots, and frequency; and (c) a big data Task Scheduling algorithm based on Thermal-aware and DVFS techniques (TSTD) to minimize the total energy consumption of data centers. The experimental results demonstrate that the proposed TSTD algorithm significantly outperforms the state-of-the-art energy efficient algorithms in terms of total energy, computing energy, cooling energy, cooling energy proportion, and total energy savings perspectives.

11 March 2025 in Computer Networks
Verified
Journal Article
Closed access
EMEhzaz Mustafa
Ehzaz Mustafa
JSJunaid Shuja
Junaid Shuja
FRFaisal Rehman
Faisal Rehman
ANAbdallah Namoun
Abdallah Namoun
Muhammad BilalMuhammad Bilal
Muhammad Bilal
KBKashif Bilal
Kashif Bilal

7 July 2017 in IEEE Access
Verified
Journal Article
Open access
KBKashif Bilal
Kashif Bilal
AEAiman Erbad
Aiman Erbad
MHMohamed Hefeeda
Mohamed Hefeeda
Advances and commoditization of media generation devices enable capturing and sharing of any special event by multiple attendees. We propose a novel system to collect individual video streams (views) captured for the same event by multiple attendees, and combine them into multi-view videos, where viewers can watch the event from various angles, taking crowdsourced media streaming to a new immersive level. The proposed system is called Cloud-based Multi-View Crowd sourced Streaming (CMVCS), and it delivers multiple views of an event to viewers at the best possible video representation based on each viewer's available bandwidth. The CMVCS is a complex system having many research challenges. In this paper, we focus on resource allocation of the CMVCS system. The objective of the study is to maximize the overall viewer satisfaction by allocating available resources to transcode views in an optimal set of representations, subject to computational and bandwidth constraints. We choose the video representation set to maximize QoE using Mixed Integer Programming. Moreover, we propose a Fairness-Based Representation Selection (FBRS) heuristic algorithm to solve the resource allocation problem efficiently. We compare our results with optimal and Top-N strategies. The simulation results demonstrate that FBRS generates near optimal results and outperforms the state-of-the-art Top-N policy, which is used by a large-scale system (Twitch).

12 April 2024 in IEEE Access
Verified
Journal Article
Open access
LJLaila Junaid
Laila Junaid
KBKashif Bilal
Kashif Bilal
JSJunaid Shuja
Junaid Shuja
ABAbdullateef Oluwagbemiga Balogun
Abdullateef Oluwagbemiga Balogun
Joel RodriguesJoel Rodrigues
Joel Rodrigues
A land administration system (LAS) is a structured framework designed to govern the management of land resources in a specific region or country. However, LAS faces challenges like inefficiencies, a lack of transparency, and susceptibility to fraud. The digitization of land records improved efficiency but failed to address manipulation, centralized databases, and double-spending issues. Traditional lease and mortgage management systems also suffer from complexity, errors, and a lack of real-time validation. At present, a significant influx of land transactions produces substantial data, classifiable as big data due to constant minute-to-minute occurrences like land transfers, acquisitions, document verification, and leasing/mortgaging transactions. In this context, we present a Blockchain-driven system that not only tackles alteration and double-spending issues in traditional systems but also implements distributed data management. Current state-of-the-art solutions do not fully incorporate crucial features of Blockchain, such as transparency, prevention of double-spending, auditability, immutability, and user participation. To tackle this problem, this research introduces a comprehensive Blockchain-powered framework for lease and mortgage management, addressing transparency, user involvement, and double-spending prevention. Unlike existing solutions, our framework integrates key Blockchain characteristics for a holistic approach. Through practical use cases involving property owners, banks, and financial institutions, we establish a secure, distributed, and transparent method for property financing. We verify the system by employing smart contracts and assess the cost and security parameters while validating the blockchain-based mortgage and lease functions.

3 May 2017 in IEEE Transactions on Cloud Computing
Verified
Journal Article
Closed access
SASahibzada Muhammad Ali
Sahibzada Muhammad Ali
MJMuhammad Jawad
Muhammad Jawad
Muhammad Usman Shahid KhanMuhammad Usman Shahid Khan
Muhammad Usman Shahid Khan
KBKashif Bilal
Kashif Bilal
JGJacob Glower
Jacob Glower
Albert ZomayaAlbert Zomaya
Albert Zomaya
Enormous energy consumption of data centers has a major impact on power systems by significantly increasing the electrical load. Due to the increase in electrical load, power systems are facing demand and supply miss-management problems. Therefore, power systems require efficient and intelligent ancillary services to maintain robustness, reliability, and stability. Data centers can provide the computational capabilities to manage power systems; however, data centers consume a tremendous amount of energy, and energy price accounts for a significant portion of their operational cost. Power system jobs will make this situation even more critical for data centers. In our work, we seek an Ancillary Services Model (ASM) to service data centers and power systems. In ASM, we find an optimal job scheduling technique for executing power systems' jobs on data centers in terms of low power consumption, reduced makespan, and fewer preempted jobs. The power systems' jobs include Optimal Power Flow (OPF) calculation, transmission line importance index, and bus importance index. Moreover, a Service Level Agreement (SLA) between data centers and power systems is shown to provide mutual benefits.