From the Rector: Covid-19

Measures by GIK against Coronavirus

GIK Institute is committed to take all necessary steps to ensure a safe environment for its students, employees and residents.
Click Here for details.

GIK Institute's measures against Coronavirus. Click here for details
From the Rector: Covid-19
Pro-Rector(Academic) message for students

FACULTY PROFILE

Engr. Dr. Sajid Anwar

Assistant Professor
Email: sajid@giki.edu.pk
Qualifications: Ph.D. Seoul National University
Research Interests: Deep Learning, Pruning and Quantizing CNN, Genetic and Evolutionary Algorithms, GPU Computing with CUDA, Embedded Systems (ARM)
Telephone: 2283

Introduction


Sajid Anwar received Ph.D. degree in electrical and computer engineering from Seoul National University, Republic of Korea, in 2017. Seoul National Univeristy is ranked 36th in the QS world university ranking. During Ph.D. research, Sajid studied the complexity of deep convolutional neural networks and proposed fixed point optimization and pruning methods. Being the pioneering work in this direction, our fixed-point optimization algorithm achieved the same level of accuracy with 2-3 bits weights rather than 32-bit weights. The second technique induce structured sparsity at all scales in a CNN. Structured pruning leads to simpler sparse representation which can be easily exploited by a generic set of modern computing platforms. My research work has received citations from the leading industrial & academic research groups like Google Deep Brain, Samsung, Microsoft, NEC, Qualcomm, Nvidia, Stanford university etc. He has been a member of faculty of computer science and engineering at Ghulam Ishaq Khan Institute of Engineering Sciences and Technology, Pakistan since 2017. 

Recently, I have focussed on the applications of deep learning algorithms. In the field of remote sensing and GIS, big data is available. This data can be utilized for understanding the fundamental variables in climate change and renewable energy. 

Research Areas: Machine learning, Deep Learning  Architectures, Genetic and evolutionary algorithms, Structured Pruning, Quantized CryptoNets, GIS & Remote Sensing, GPU Computing with CUDA, ARM Assembly Level Optimization, Video Decoder (MPEG4 Simple Profile)

Selected Publications


  • Anwar Sajid, Kyuyeon Hwang, and Wonyong Sung. "Structured Pruning of Deep Anwar S, Hwang K, Sung W. Structured pruning of deep convolutional neural networks. ACM Journal on Emerging Technologies in Computing Systems (JETC). 2017 Feb 9;13(3):32.
  • Hussain, Farhan, Mian Ilyas Ahmad, Sajid Anwar, Aimal Khan, and Pyoung Won Kim. "Efficient motion estimation using two-bit transform and modified multilevel successive elimination." Journal of Ambient Intelligence and Humanized Computing(2018): 1-7.
  • Ahmad, Maqbool, Khan Alam, Shahina Tariq, Sajid Anwar, Jawad Nasir, and Muhammad Mansha. "Estimating fine particulate concentration using a combined approach of linear regression and artificial neural network." Atmospheric Environment (2019): 117050.
  • Halim, Z., Khan, A., Sulaiman, M., Anwar, S. and Nawaz, M., On finding optimum commuting path in a road network: A computational approach for smart city traveling. Transactions on Emerging Telecommunications Technologies.
  • Anwar, Sajid, and Wonyong Sung. "Compact deep convolutional neural networks with coarse pruning." arXiv preprint arXiv:1610.09639 (2016).

 

  • Anwar Sajid, Kyuyeon Hwang, and Wonyong Sung. "Fixed-point optimization of deep convolutional neural networks for object recognition." In Acoustics, Speech and Signal Processing (ICASSP), 2015 IEEE International Conference on, pp. 1131-1135. IEEE, 2015.
  • Anwar Sajid, Kyuyeon Hwang, and Wonyong Sung. “Learning separable fixed point kernels with deep CNN.” In Acoustics, Speech and Signal Processing (ICASSP), 2016 IEEE International Conference on, IEEE, 2016.
  • Anwar Sajid and Wonyong Sung, “GPU Based Digital Signal Processing Filtering” KICS Korea, 2013
  • Anwar Sajid and Wonyong Sung, “GPU Based Digital Signal Processing Filtering” IEEK 2009
  • Anwar Sajid and Wonyong Sung, “OpenMax Based MPEG4 SP Video Decoder for ARM926EJ-S Platform” IEEK 2009

 

Research Projects


  1. Characterization of Alzheimer’s disease by classification of subcortical regions using deep learning framework, Co-PI (1.1 Million PKR)
  2. Ask Lephrechaun (2nd Runner-up at NGIRI National level Competition)
  3. Intelligent Mining System (Autonomous Tunnel Mapping)

Open Source Simulators


The following repository hosts the implementation of convolutional neural network.

Both CPU and GPU versions are provided.

https://github.com/anwarsajid

 

Teaching


Undergraduated Courses

  • Artificial Intelligence
  • Deep Learning (Graduate course)
  • Software Engineering-I
  • Introduction to Computing (C++ fundamentals)
  • Real-Time Embedded Systems
  • Introduction to Soft Computing
  • Design and Analysis of Algorithms

Final Year Projects


2019

  • The SketchArtist with Generative Adversarial Networks (GAN)
  • Turon: The AI Medical Assistant
  • Resource Pooling for Courier Service
  • AI Imagination

2018

  • Tunnel Mapping with UAV
  • Autonomous Driving with Toy Cars (inspired from DuckieTown)
  • Quantized Fully Convolutional Neural Networks
  • Attendance via Gait Analysis

2017

  • Intelligent Bionic Hand (sEMG signals, Deep neural network, Arduino, 3D printing)
  • Ask Leprechaun (RNN based Crypto-Currency trend prediction, 1st prize holder)

Memberships


  1. Member Pakistan Engineering Coucil 
  2. Member BoS UET Mardan (Computer Systems Engineering)
  3. Member BoS UAJK (Software Engineering)

Startups


Research Students


  • Mr. Jalees Ur Rahman (Optical Flow Based Classification with Deep Learning)
  • Mr. Usman Haider (Efficient Inference using Pruned ConvNets)
  • Mr. Muhammad Asif (Hierarchichal Clustering of Confusion Matrix)
  • Mr. Musavir Ali (Quantization of Deep Learning Algorithms)
  • Mr. Muhammad Rashid (Deep Learning Applications in Remote Sensing and GIS)


Edit Profile