HBKU CSE Student Develops Winning Propaganda Detection Model

College of Science and Engineering Student Develops Winning Propaganda Detection Model

18 Nov 2019

Challenge as part of 2019 Workshop on Natural Language Processing for Information Freedom

Shehel Yoosuf, a Masters in Data Science and Engineering student in Hong Kong.

A student from Hamad Bin Khalifa University’s (HBKU) College of Science and Engineering (CSE) was among the winners at the 2019 workshop on NLP4IF co-located with Empirical Methods in Natural Language Processing (EMNLP), held in Hong Kong from November 3 to 7. 

Shehel Yoosuf, a Masters in Data Science and Engineering student working under the guidance of Dr. David Yang, leading a QNRF-funded project on Fraudulent Message Detection Under Differential Privacy, developed the best performing system in the AI-driven fine grained propaganda detection task. More than 15 international teams over a period of 2 months worked on developing Machine Learning models capable of finding every instance of propaganda from a set of news articles. The participant’s system was required to specify the phrase where propaganda occurs and also determine the type of propaganda utilized from a list of 18 techniques. Categories for the latter task included ‘appeal to authority’, ‘bandwagons’ and ‘black and white fallacy.’

Speaking after receiving the award, Shehel said: “It was an honor to represent HBKU in the competition and present the solution in the workshop. I enjoyed the challenge of working on such a relevant problem as propaganda which is becoming widespread with the prevalence of web media. I’m grateful to CSE for providing me with this opportunity and resources to make it happen.”

Speaking after the event Dr. Mounir Hamdi, Dean of CSE, said: “We’re naturally delighted that one of our students excelled in the conference’s propaganda detection competition. The standard of contestants was high, as were their abilities to cope with the demands of this task. Shehel’s success underscores CSE’s growing reputation as a world-class place to study science, engineering and technology. It also bodes well for our mission to advance knowledge and nurture future leaders and innovators through teaching and research across a range of carefully targeted programs.”

This year’s propaganda detection competition was organized by HBKU’s Qatar Computing Research Institute and the University of Bologna. 

CSE regularly holds events to highlight its research activities and projects. For more information, please visit cse.hbku.edu.qa