“I've worked under the direct & indirect supervision of Hussein for 3 years at Mawdoo3, In those 3 years we have worked on multiple research projects and Salma, the Arabic Virtual Assistant. Hussein technical proficiency has always been crucial to the projects he managed closely, continuously and constantly providing insightful remarks and recommendations that would critically impact the project's output or save us a great amount of time. Hussein has a rare mixture of product management, in-depth understanding of the technologies utilized and great people skills that made him essential throughout any project's life-cycle, either by aligning the product's vision with the technological capabilities and limitations, resolving tough conflicts across teams .. and the list goes on. Hussein managed to build a culture of innovation, excellence and loyalty at Mawdoo3, although I would only usually speak for myself, but I will not be exaggerating if I said that everyone at the AI department felt privileged to have Hussein as our leader.”
Hussein AL-NATSHEH
Abu Dhabi, Abu Dhabi Emirate, United Arab Emirates
17K followers
500+ connections
About
Driven by my fascination for cutting-edge AI technology and its transformative power…
Experience
Education
-
Université de Lyon
-
Activities and Societies: PhD Student Representative at the Counsel of Doctoral School of the University of Lyon, France
Data science with a thesis of using machine learning in text mining and automatic text understanding for information retrieval systems
L'Ecole Doctorale InfoMaths (University de Lyon)
ERIC laboratory (https://eric.msh-lse.fr/en/presentation/) (Unité de Recherche des Universités Lyon 2 et Lyon 1)
CNRS
A full scholarship by the French government (Regional scientific research fund) -
-
-
-
A double master degrees (2 universities) of 2-year specialized courses including 6-month practical internship. The program is a collaboration between 6 European universities providing their best courses in Data Mining and Knowledge Management. Out of more than 800 of applicants, only 18 students were carefully selected for 2013-2015.
-
-
Activities and Societies: IEEE Jordan SAC
Master Thesis in the field of Machine Learning and Data Mining
-
-
Activities and Societies: IEEE Student Branch Chairman
Computer Engineering student
Chair and co-founder of IEEE student branch
Licenses & Certifications
Volunteer Experience
-
Program Committee Member
Arabic Natural Language Processing Workshop (WANLP)
- Present 4 years 8 months
Science and Technology
Committee member of the fifth (2020) and the sixth version (2021)
-
PC Member
ICNLSP 2021
- Present 3 years 9 months
Science and Technology
-
Board Member
Unihance
- Present 4 years 2 months
Education
Started as a mentor of the founder, then an advisor of the startup. Currently, an investor and a board member.
-
Program Committee Member
ACLing 2021
- Present 4 years 3 months
Science and Technology
-
Arabic Teacher at Badr (Children School of CCMPG)
Centre Culturel des Musulmans du Pays de Gex (CCMPG)
- 5 months
Education
-
PhD Students Representative
InfoMaths Doctoral School of the University of Lyon
- 2 years
Education
Publications
-
Deep Contextualized Pairwise Semantic Similarity for Arabic Language Questions
https://arxiv.org
Question semantic similarity is a challenging and active research problem that is very useful in many NLP applications, such as detecting duplicate questions in community question answering platforms such as Quora. Arabic is considered to be an under-resourced language, has many dialects, and rich in morphology. Combined together, these challenges make identifying semantically similar questions in Arabic even more difficult. In this paper, we introduce a novel approach to tackle this problem…
Question semantic similarity is a challenging and active research problem that is very useful in many NLP applications, such as detecting duplicate questions in community question answering platforms such as Quora. Arabic is considered to be an under-resourced language, has many dialects, and rich in morphology. Combined together, these challenges make identifying semantically similar questions in Arabic even more difficult. In this paper, we introduce a novel approach to tackle this problem, and test it on two benchmarks; one for Modern Standard Arabic (MSA), and another for the 24 major Arabic dialects. We are able to show that our new system outperforms state-of-the-art approaches by achieving 93% F1-score on the MSA benchmark and 82% on the dialectical one. This is achieved by utilizing contextualized word representations (ELMo embeddings) trained on a text corpus containing MSA and dialectic sentences. This in combination with a pairwise fine-grained similarity layer, helps our question-to-question similarity model to generalize predictions on different dialects while being trained only on question-to-question MSA data
-
NSURL-2019 Shared Task 8: Semantic Question Similarity in Arabic
https://arxiv.org
Question semantic similarity (Q2Q) is a challenging task that is very useful in many NLP applications, such as detecting duplicate questions and question answering systems. In this paper, we present the results and findings of the shared task (Semantic Question Similarity in Arabic). The task was organized as part of the first workshop on NLP Solutions for Under Resourced Languages (NSURL 2019) The goal of the task is to predict whether two questions are semantically similar or not, even if…
Question semantic similarity (Q2Q) is a challenging task that is very useful in many NLP applications, such as detecting duplicate questions and question answering systems. In this paper, we present the results and findings of the shared task (Semantic Question Similarity in Arabic). The task was organized as part of the first workshop on NLP Solutions for Under Resourced Languages (NSURL 2019) The goal of the task is to predict whether two questions are semantically similar or not, even if they are phrased differently. A total of 9 teams participated in the task. The datasets created for this task are made publicly available to support further research on Arabic Q2Q.
-
Mawdoo3 AI at MADAR Shared Task: Arabic Tweet Dialect Identification
ACL
Arabic dialect identification is an inherently complex problem, as Arabic dialect taxonomy is convoluted and aims to dissect a continuous space rather than a discrete one. In this work, we present machine and deep learning approaches to predict 21 fine-grained dialects form a set of given tweets per user. We adopted numerous feature extraction methods most of which showed improvement in the final model, such as word embedding, Tf-idf, and other tweet features. Our results show that a simple…
Arabic dialect identification is an inherently complex problem, as Arabic dialect taxonomy is convoluted and aims to dissect a continuous space rather than a discrete one. In this work, we present machine and deep learning approaches to predict 21 fine-grained dialects form a set of given tweets per user. We adopted numerous feature extraction methods most of which showed improvement in the final model, such as word embedding, Tf-idf, and other tweet features. Our results show that a simple LinearSVC can outperform any complex deep learning model given a set of curated features. With a relatively complex user voting mechanism, we were able to achieve a Macro-Averaged F1-score of 71.84% on MADAR shared subtask-2. Our best submitted model ranked second out of all participating teams.
-
Metadata Enrichment of Multi-disciplinary Digital Library: A Semantic-Based Approach
Springer International Publishing
In the scientific digital libraries, some papers from different research communities can be described by community-dependent keywords even if they share a semantically similar topic. Articles that are not tagged with enough keyword variations are poorly indexed in any information retrieval system which limits potentially fruitful exchanges between scientific disciplines. In this paper, we introduce a novel experimentally designed pipeline for multi-label semantic-based tagging developed for…
In the scientific digital libraries, some papers from different research communities can be described by community-dependent keywords even if they share a semantically similar topic. Articles that are not tagged with enough keyword variations are poorly indexed in any information retrieval system which limits potentially fruitful exchanges between scientific disciplines. In this paper, we introduce a novel experimentally designed pipeline for multi-label semantic-based tagging developed for open-access metadata digital libraries. The approach starts by learning from a standard scientific categorization and a sample of topic tagged articles to find semantically relevant articles and enrich its metadata accordingly. Our proposed pipeline aims to enable researchers reaching articles from various disciplines that tend to use different terminologies. It allows retrieving semantically relevant articles given a limited known variation of search terms. In addition to achieving an accuracy that is higher than an expanded query based method using a topic synonym set extracted from a semantic network, our experiments also show a higher computational scalability versus other comparable techniques. We created a new benchmark extracted from the open-access metadata of a scientific digital library and published it along with the experiment code to allow further research in the topic.
-
Étiquetage thématique automatisé de corpus par représentation sémantique
EGC 2018, vol. RNTI-E-34, pp.323-328
-
UdL at SemEval-2017 Task 1: Semantic Textual Similarity Estimation of English Sentence Pairs Using Regression Model over Pairwise Features
Association for Computational Linguistics
-
Commercializing Computational Intelligence Techniques in a Business Intelligence Application
IEEE
This paper reports on the commercialization of a business intelligence application deploying computational intelligence techniques. Theoretical foundations are included where appropriate, along with implementation results and comparative benchmarks...
Other authorsSee publication -
Performance optimization of adaptive resonance neural networks using genetic algorithms.
IEEE Xplore Press, Foundations of Computational Intelligence, 2007. FOCI 2007
We present a hybrid clustering system that is based on the adaptive resonance theory 1 (ART1) artificial neural network (ANN) with a genetic algorithm (GA) optimizer, to improve the ART1 ANN settings. As a case study, we will consider text clustering. The core of our experiments will be the quality of clustering, multi-dimensional domain space of ART1 design parameters has many possible combinations of values that yield high clustering quality. These design parameters are hard to estimate…
We present a hybrid clustering system that is based on the adaptive resonance theory 1 (ART1) artificial neural network (ANN) with a genetic algorithm (GA) optimizer, to improve the ART1 ANN settings. As a case study, we will consider text clustering. The core of our experiments will be the quality of clustering, multi-dimensional domain space of ART1 design parameters has many possible combinations of values that yield high clustering quality. These design parameters are hard to estimate manually. We proposed GA to find some of these sets. Results show better clustering and simpler quality estimator when compared with the existing techniques. We call this algorithm genetically engineered parameters ART1 or ARTgep
Other authorsSee publication
Courses
-
Advanced Business Internship in Technoport / Luxembourg
-
-
Advanced Databases
-
-
Complex Data Warehousing
-
-
Data Processing: Cleaning, feature selection, feature construction
-
-
Empretec
-
-
Facilitation Skills Workshop by USAID
-
-
Intellectual Property Rights (IPR) Management, Licencing and Technology Transfer
-
-
Logic and Knowledge Representation
-
-
Machine Learning
-
-
Methodology and Tools for Research
-
-
Mining Complex Data: Text, Image, Web
-
-
Modelling Complex Systems in Social Science
-
-
Multidimensional Data Analysis
-
-
Optimization
-
-
Probability and Statistics
-
-
Project Management, HRM, Marketing, Sales, Business Planning
-
-
Software Methodologies
-
-
Symbolic Learning
-
-
TRIZ and Systematic Innovation in Business and Technology
-
Projects
-
Aramco MetaBrain Industrial LLM
On-prem fine-tuned LLM on Aramco proprietary documents and a RAG conversational AI applications on top of it
-
Aramco Downstream Global Optimizer
Data and AI Technology Lead/Executive
-
Beyond Search
Enterprise data LLM powered semantic search platform. It enables companies to index their proprietary data before using fine tuned LLM for conversational question answering for more trusted and referenced answers
-
Blend Optimizer
A no-code machine learning and Optimization SaaS products for regulated chemical industries including for example, Lubrication, gasoline blending, non-metalic, and pharmaceutical products.
-
Smart Gates
A fully dynamic workflow builder for Autonomous Border Control (ABC). It replaces sensor based systems with a machine vision systems that better capture fraudulent cases.
-
Author Name Disambiguation
-
Author name disambiguation based on distances between the citation details of the publications. The data model implements co-reference and co-authorship graphs as well as hierarchical clustering with a dynamic detection of the number of clusters. Experiments and results show superiority over modularity and community graph based clustering.
Other creators -
-
OneCard business Intelligence and high-delivery-rate targeted marketing email program
-
Implementation of a self-service online business intelligence and data mining portal for the marketing and managerial team linked with a targeted email marketing system. The email program provide very high inbox delivery rate following anti-spam best practices. Targeting is based on not only calculated demographics but also behavioral predictive analysis of the online users. The business intelligence service also sends periodic emails listing some important visualized trending reports and KPIs…
Implementation of a self-service online business intelligence and data mining portal for the marketing and managerial team linked with a targeted email marketing system. The email program provide very high inbox delivery rate following anti-spam best practices. Targeting is based on not only calculated demographics but also behavioral predictive analysis of the online users. The business intelligence service also sends periodic emails listing some important visualized trending reports and KPIs to OneCard.net for better market understanding and business decisions.
Other creators -
Personalized Recommender System for ChoozOn Corp (BlueKangaroo.com)
-
Real-time items recommendation for each user based on his profile, similar users and social activities. The system also infers some users interests and utilize that with the user's preferences to predict a list of items he would likely buy.
Other creatorsSee project -
Cloud based Cross-Selling and Churn Management Solution for FoodCity Supermarkets(Piloting)
-
By analysing the users' spending amounts from each supermarket department, the system predict the possibility of loosing each loyal customer based on his similarity to users who have recently cut their spending from the supermarket.
-
Research-Industry Jordanian Applied Research Professionals Networking Hub Software-as-a-Service Solution for IPCO
-
-
Data Warehousing and ERP (LIMS) project implementation of Ministry of Water and Irrigation in Jordan
-
-
Cross-Selling and Customer Retention Data Mining Application for Telecom (AutoVAS)
-
behavioural targeting system built on customer profiler (clustering) engine that predicts relevant potential buyers based on their historical transactions.
Other creatorsSee project
Honors & Awards
-
Gold Award
Huawei Developer Competition 5 Countries
Mowjaz App team ranked first in this competition of best Huawei AppGallary App
-
PhD Scholarship
CNRS
-
Full Scholarship from Erasmus Mundus (EM DMKM)
-
Full Scholarship from Erasmus Mundus master course in Data Mining and Knowledge Management (EM DMKM) a project funded by the European Union
-
Top 30 Arab Tech Innovators Under 30
UMEN Magazine
Named as one of the "Top 30 Arab Tech Innovators Under 30" . Ranked 18 by Fouad Jeryes in UMEN Magazine
-
Semi-Finalist
MIT Arab Business Plan Competition
-
Top 50 SMEs
InfoDev, 4th Global Forum Innovation and Technology Entrepreneurship
Selected among the top 50 SMEs by InfoDev, 4th Global Forum Innovation and Technology Entrepreneurship, Helsinki, Finland
-
Arab GoldenChip Award : Top 3 Best Software for Export
MENA ICT Week, Bahrain
-
The best Jordanian start-up
MedVentures Award
MedVentures Award 2010, Marseille, FRANCE
-
First Award in Arab Technology Business Plan Competition
Arab Science and Technology Foundaion
First Award in Arab Technology Business Plan Competition - Seed Stage Category 2008-2009
-
Dedication Award
First IEEE Middle East Student Branch Congress
-
Third Award
Queen Rania National Entrepreneurship Competition
-
Best paper award at KESW 2016
The 7th International Conference on Knowledge Engineering and Semantic Web, Prague, 2016
Languages
-
English
Full professional proficiency
-
Arabic
Native or bilingual proficiency
-
French
Limited working proficiency
Organizations
-
Jodan Engineers Association
-
- Present
Recommendations received
11 people have recommended Hussein
Join now to viewOther similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore MoreOthers named Hussein AL-NATSHEH
1 other named Hussein AL-NATSHEH is on LinkedIn
See others named Hussein AL-NATSHEH