Hadi Abdi Khojasteh
Hadi Abdi Khojasteh
Computer Science Researcher
hkhojasteh at iasbs dot ac dot ir

BioI am an R&D Software Architect of the Mehad Sanat Incorporation and Researcher at the Institute for Advanced Studies in Basic Sciences (IASBS). I have been designed, analyzed and implemented several high-tech automotive industry equipment, car diagnostic tools and real-time software systems, cooperated with a team of researchers and professional engineers.

I received my BEng degree in information technology engineering form IASBS in 2016 and was a graduate of computer science at the same institute and studies the behaviour of Human-Centred Artificial Intelligence Systems in vision and natural language aspects, where was advised by Dr Ebrahim Ansari and Dr Parvin Razzaghi. I also participated in the European Union project consists of Vienna University of Technology, University of Bonn, Aristotle University of Thessaloniki and two European SMEs focused on Data Science. I have skill in Machine Learning, Deep Learning, Computer Vision, Natural Language Processing, High-performance computing and Embedded programming. In my free time, I am mountaineering, backpacking, hiking and blogging.


Education

Scholar, PI Campus, October – December 2019

Pi School, School of Artificial Intelligence (AI), Rome, Italy

The School of AI hosts a batch of the best researchers around the world. Each of the participants receives personalised guidance from an expert on real industry projects. Advisors are from top universities such as MIT, Carnegie Mellon University, University of Cambridge and University of Sussex and world-leading tech companies such as Google, Facebook, Amazon and fast-growing startups in this programme. I contributed to a project presented by Octo Telematics, developing AI-based business solutions.


Master of Science (MSc), Computer Science, 2016 – 2019

Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan, Iran

Thesis title: "Automatic Image Description Generation Using Deep Multimodal Embeddings"
DOI: 10.13140/RG.2.2.21497.01129, Supervisors: Dr Ebrahim Ansari and Dr Parvin Razzaghi
Research on Deep Learning, Computer Vision and Natural Language Processing.

GPA: 3.73/4.00 (Ranked 1st) via 34 passed credit hours

Selected courses passed: Advanced Artificial Intelligence (A+), Parallel Algorithms (A+), Formal Methods (A+), Machine Learning (A+), Computer Vision (A), Information Retrieval (A+), Logic Programming (A+)

Visiting Researcher, Aristotle University of Thessaloniki (AUTH), April - May 2019

School of Informatics, Faculty of Sciences, Aristotle University of Thessaloniki (AUTH), Thessaloniki, Greece

Studied as an EU funded Erasmus+ project called GraDAna in partnership with Vienna University of Technology (TUVIE), University of Bonn (UoB), Aristotle University of Thessaloniki (AUTH) and 9 other partners and universities.

Participated topics: Statistical Methods in Information Systems, Open Source Software Data Engineering, Social Networks Analytics, Cloud Computing, Distributed Systems, Citizen Science and Smart Cities, Big Data, Big Data in Practice, Tools and Technologies in Data Science, Upgrading Soft Skills for Computer Engineers, Conducting "Open Science": Approaches, Tools and Practices


Bachelor of Engineering (BEng), Information Technology Engineering, 2012 – 2016

Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan, Iran

Thesis title: "A Survey on patch-based synthesis: GPU Implementation and Optimization"
DOI: 10.13140/RG.2.2.29490.86729/1, Supervisor: Dr Ebrahim Ansari

GPA: 3.36/4.00 (Ranked 5th) via 210 passed credit hours

Selected courses passed: Advanced Programming (A+), Discrete Mathematics (A+), Data Structures (A+), Statistics and Probability (A+), Automata Theory (A+), Management (A+), Digital Electronics (A+), Artificial Intelligence (A), Computer Networks (A), Neural Networks (A+), Multimedia (A+), Computer Networks (A+), Databases (A), English (A+), Human-Computer Interaction (A)

Human-Centered Semi-Autonomous Vehicle

I am a member of the Human-Centered Semi-Autonomous Vehicle project (known as FARAZ), which is an academia and industry collaboration that aims to create a semi-autonomous car by integrating some state-of-the-art approaches in computer vision and machine learning for assisting the drivers during critical and risky moments in which driver would be unable to steer the vehicle safely.

Our collaborators include Institute for Advanced Studies in Basic Sciences (IASBS), Charles University in Prague, Institute for Research in Fundamental Sciences (IPM), and Mehad Sanat Incorporation.

Learn More
Our Team. Left to right: Hadi Abdi Khojasteh, Ebrahim Ansari, Parvin Razzaghi and Alireza Abbas Alipour - Faraz the Human-Centered Semi-Autonomous Vehicle

Publications

Enhanced Large Scale Colloquial Persian Language Understanding
Enhanced Large Scale Colloquial Persian Language Understanding

Hadi Abdi Khojasteh, Ebrahim Ansari, Mahdi Bohlouli, Accepted
Proceedings of the 12th International Conference on Language Resources and Evaluation (LREC), 2020

Language recognition has been significantly advanced in recent years by means of modern machine learning methods such as deep learning and benchmarks with rich annotations. This consists of a significant gap in describing the colloquial language especially for low-resourced ones such as Persian. We propose a "Large Scale Colloquial Persian Dataset" (LSCP). LSCP is hierarchically organized in a semantic taxonomy that focuses on multi-task informal Persian language understanding as a comprehensive problem. The proposed corpus consists of 120M sentences resulted from 27M tweets annotated with parsing tree, part-of-speech tags, sentiment polarity and translation in five different languages. The LSCP includes 120M sentences from 27M casual Persian tweets with its derivation tree, part-of-speech tags, sentiment polarity and translations in English, German, Czech, Italian and Hindi spoken languages.

Website Paper Data
Deep Multimodal Image-Text Embeddings for Automatic Cross-Media Retrieval
Deep Multimodal Image-Text Embeddings for Automatic Cross-Media Retrieval

Hadi Abdi Khojasteh, Ebrahim Ansari, Parvin Razzaghi, Akbar Karimi, In Preparation

This study considers the task of matching images and sentences by learning a visual-textual embedding space for cross-modal retrieval. We introduce an end-to-end deep multimodal convolutional-recurrent network for learning both vision and language representations simultaneously to infer image-text similarity. To learn about the joint representations, we leverage our newly extracted collection of tweets from Twitter. The main characteristic of our dataset is that the images and tweets are not standardized the same as the benchmarks. Experimental results on MS-COCO benchmark dataset show that our model outperforms certain methods presented previously and has competitive performance compared to the state-of-the-art.

Paper
An Intelligent Safety System for Human-Centered Semi-Autonomous Vehicles
An Intelligent Safety System for Human-Centered Semi-Autonomous Vehicles

Hadi Abdi Khojasteh, Alireza Abbas Alipour, Ebrahim Ansari, Parvin Razzaghi, Published
Oral presentation in the international conference on Contemporary issues in Data Science (CiDaS), 2018
Published as a chapter in Data Science: From Research to Application Springer book, 2020

Monitoring drivers actions by computer vision techniques to detect driving mistakes in real-time and then planning for autonomous driving to avoid vehicle collisions is one of the most important issues that has been investigated in the machine vision and Intelligent Transportation Systems. To avoid driving incidents, this paper proposes an integrated safety system that continuously monitors the drivers attention and vehicle surroundings, and finally decides whether the actual steering control status is safe or not.

Website Paper

Professional Experiences

R&D Software Architect

Mehad Sanat Inc. (founded in 1996)
October 2014 – Present, Karaj, Iran

Designed and implemented several high-tech automotive industry equipment, car diagnostic tools and real-time software systems, cooperated with a team of 17 researchers and professional engineers.

My responsibilities included:
- Analysis, design and development of software systems
- Analysis, design and development of data transmission over dedicated networks
- Software programming for proprietary hardware
- Big data analytics and visualizations
- Embedded programming
- Design and development of custom web servers and services


University Lecturer

Department of Computer Science of Institute for Advanced Studies in Basic Sciences (IASBS)
September 2019 – Present, Zanjan, Iran

Taught and developed campus-based courses for university modules. I contributed to three courses about 18 credit hours for three semesters and designed the following courses:

Sep 2019 – Jul 2020
- Information Technology Engineering (9 credit hours based on 3 courses)
Jan 2019 – Jul 2020
- Advanced Computer Programming (6 credit hours based on 2 courses)
Sep 2019 – April 2020
- Software Engineering (3 credit hours based on 1 course)

Teaching Assistant

Department of Computer Science of Institute for Advanced Studies in Basic Sciences (IASBS)
September 2013 – September 2019, Zanjan, Iran

Assisted over 350 graduate and undergraduate students, taught courses supplementary materials, and evaluated projects, presentations and quizzes.
I have been involved in following courses:

April – Jul 2019
- Text Mining and Web Mining (Graduate Course) Instructed by Dr Ebrahim Ansari and Dr Mahdi Bohlouli
Sep – Dec 2018
- Advanced Algorithms (Graduate Course) Instructed by Dr Mansoor Davoodi Monfared and Dr Ebrahim Ansari
April – Jul 2018
- Information Retrieval (Graduate Course) Instructed by Dr Ebrahim Ansari
Jan - April 2018
- Machine Learning (Graduate Course) Instructed by Dr Parvin Razzaghi
Sep – Dec 2017
- Computer Programming Basics Instructed by Dr Ebrahim Ansari
April – Jul 2017
- Intelligent System Design Instructed by Dr Parvin Razzaghi
Dec 2016 – Jul 2017
- Advanced Computer Programming Instructed by Dr Ebrahim Ansari
Oct 2016 – Jan 2017
- Advanced Data Structures Instructed by Dr Ali Khavasi
Oct 2015 – Jan 2016
- Artificial Intelligence Instructed by Dr Parvin Razzaghi
Dec 2014 – Jul 2015
- Advanced Computer Programming Instructed by Dr Ebrahim Ansari
Jan - April 2014
- Advanced Computer Programming Instructed by Dr Mahmoud Shirazi
Sep – Dec 2013
- Computer Programming Basics Instructed by Dr Rozita Jamili

Pet Projects

Break a Golestan CAPTCHA system with Machine Learning

Break a Golestan CAPTCHA system with Machine Learning

How to break a CAPTCHA system with Machine Learning is my attempt at explaining Convloutinal Neural Networks with empirical case study, relying more on code and intuitions than mathematics. In this study, I intended to break the Golestan system's CAPTCHA using the machine/deep learning and statistical methods in an educational step-by-step approach.

Interactive Web Spider

Interactive Web Spider

The goal of the project was to create a web crawler for getting and categorizing web pages. After reviewing each page of the site, crawler extract new URLs and get new pages. The created crawler is resistant to malicious pages, duplicate pages and pages outside the domain.

Alongside the computing core, we have built a panel to show the function of the entire system, which displays the information as a large graph. This platform was implemented in Python.

Bag of Visual Words (BoVW) for Image Classification

Bag of Visual Words (BoVW) for Image Classification

The program will generate a visual vocabulary and train a classifier using a provided set of already classified images. We will examine the task of image classification starting with making visual word vocabulary by generating a codebook of visual words and then move on to techniques that resemble the aggregating the histograms of the visual words and linear support vector machine (SVM) classifier learned by histograms. The Open Source C++ implementation available on Github.

Skills and Expertise

C
STL, Boost, OpenMPI, OpenMP, pthreads, OpenGL, FFmpeg, libgd, libcurl, LAPACK, CMake, Doxygen, GCC
C++
STL, Boost, Qt, CUDA, TensorFlow, OpenCV, OpenMP, FLTK, OpenSceneGraph, GD, Armadillo, OpenBLAS, libcurl, Torch, Cython, CMake, Doxygen, GCC
.NET
C#, ASP.NET MVC, .Net Core, Identity, Entity Framework, Xamarin, SQL Connector, dotnet-service, Web Socket, WPF, OpenCvSharp, NuGet, MSBuild
JAVA
Native, Groovy, Apache Maven, Apache Mahout, Apache Spark, Elasticsearch
Python
NumPy, Pandas, OpenCV, Multiprocessing, Keras, TensorFlow, scikit-learn, Theano, SciPy, scikit-learn, Matplotlib, Bokeh, NLTK, polyglot, spacy, BeautifulSoup, Cython, IPython, MicroPython
Embedded
FreeRTOS, Embedded C, ARM Microcontrollers
PHP
Native, Laravel, Yii
Web
HTML, CSS, JavaScript, jQuery, WebGL
Graphics
Adobe Photoshop, Adobe Lightroom, Adobe Illustrator
Others
Erlang, Prolog, Matlab, Altium Designer, Lua, SQL, LaTeX, Git, SVN

Selected Research Areas: Computer vision, natural language processing, pattern recognition, machine learning (deep learning methods), artificial neural networks, image and signal processing (for object recognition/detection, segmentation, scene and video analysis), algorithms, data structures, programming languages, parallel and distributed computing (for GPU and CPU clusters/grid), big data analytics, data science, also familiar with probability and statistics, linear algebra, discrete mathematics, statistical data analysis, neuroscience and practical electronics.

Gallery