I am currently a masters student at the Computing Science Faculty of the University of Alberta, expecting to graduate on September, 2018. My main research interests are Machine Learning, Data Science and Artificial Intelligence.
I did my undergraduate studeis in Electrical Engineering at the University of Tehran , the most prestigious university in Iran.
I was born in Yazd, a beautiful city in Iran, on October 17th, 1992. I received my diploma in Physics and Mathematics discipline from Shahid Sadoughi High School, under the supervision of NODET (National Organization for Developing Exceptional Talents).
School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran, 2016.
Total GPA :
Shahid Sadoughi High School, under the supervision of NODET (National
Organization for Developing Exceptional Talents), Yazd, Iran.
GPA : 19.63/20 (4/4)
Machine Learning (Instructor: Prof. Russel Greiner, Final Grade: 4/4)
Reinforcement Learning (Instructor: Prof. Richard Sutton, Final Grade: 4/4)
Deep Learning (Instructor: Prof. Dale Schuurmans)
Data Mining in Rich Data (Instructor: Prof. Osmar Zaiane)
Probabilistic Graphical Models (Instructor: Prof. Russel Greiner)
Pattern Recognition (Graduate Course)
Data Structure and Algorithm
Probability and Statistics Engineerig
Signals and Systems
Digital Signal Processing (Graduate Course)
Digital Logic Design
Introduction to Biomedical Engineering
Physiology 1 and 2
Introduction to Medical Physics
Introduction to Foundations of Computation (University of Alberta, Instructor: Prof. Osmar Zaiane, Winter 2017)
Introduction to Foundations of Computation (University of Alberta, Instructor: Prof. Greg Kondrak, Fall 2016)
Engineering Mathematics (Uinversity of Tehran, Instructor: Prof. Mahmoud Mohammad-Taheri, Spring 2015)
Microprocessors (University of Tehran, Instructor: Dr. Omid Fatemi, Spring 2014, Fall 2014, Spring 2015)
Radiology Systems (University of Tehran, Instructor: Prof. Hamid Soltanian-Zadeh, Fall 2015)
IoT or Internet of Things is a scenario in which devices around you can send data over a network without your direct involvement. When you use Intelligent Dumbbell for lifting weights, this device counts the number of your lifts using gyroscope module and sends out data to our website database uisng WiFi module. The user can send data to network using specific hand movements such as twisting his wrist in counter-clockwise diection. The website was designed using Ruby on Rails. You can find more information about Intelligent Dumbbell and a little presentation of this project Here.
Fall detection is a major concern in many appliactions including rehabilitation methods for detecting fall for elderly people. In this project we designed a way to detect fall. Images of a specific scene are captured using a digital camera and then processed for silhouette extraction. After image processing part we extract features from the subject's silhouette and decide whether the subject has fallen or not by feeding these data into a SVM classifier we trained earlier. For more information you can see a report of this project in pdf format or a detailed powerpoint representaion of our project.
In the picture you see a reindeer image and the result of segmentation based on color of the picture. There are many methods for image segmentation and here is the Mixture Guassian Micture Model method for segmentation of image using its RGB features. For more information you can see a brief report or get the MATLAB code.
Here's a little agent for solving Persian Dooz game. It's implemented using minimax algorithm with alpha-beta pruning. You can download the code here and compete with it (Of course with your agent :) )
Nowadays with the advance of digital technology there has been a huge growth in the amount of data which have been created in different fields of science. Many of applications in different fields of science exploit the use of classifiers such as SVM, which gives reasonable results when you have a huge amount of data, and other classifiers for classifying information and input data and then decide based on the classification results. The problem here is that many of the data we use have huge dimensionality (many features of data are considered) and also some of these features carry the same information and have correlation with each other. Feature Conditioning tackles with this problem and tries to find the best features of data which carry the most information possible and have little correlation. In this project we tried some the best known feature conditioning methods such as PCA, LDA, Kernel PCA and LLE on synthetic data and then we feed the feature conditioned(!) data into a few classifiers such as SVM, KNN and Bayes optimal classifier to assess these methods. For more information you can read our detailed report (In persian language (!)) of this project.
In this project we implemented a simple voice recognition system for recognizing the user's voice (for simplicity the system was designed to discern user's voice saying numbers between 1 to 10) based on MFCC (Mel frequency cepstral coefficients) features of user's voice. The processed voice is then feed into a Neural Network, which was trained already using user's voice, for classification. For more information read the detailed report.
If you need to contact me you can use these links and addresses. I've also provided my Social Network links .