Princess Sumaya University for Technology has witnessed its first defense of a PhD dissertation in the Computer Science program. The dissertation was presented by student Wael Al-Ataiwi under the title: "Summarizing Arabic Texts Using Machine Learning of Meaning Networks". The defense of the dissertation, which is the first of its kind in processing and extracting abstracts of texts in Arabic, took place at PSUT under the supervision of Prof. Arafat Awajan.
PSUT’s President, Prof. Mashhoor Al-Refai, expressed his great pleasure that this dissertation was the first fruit of this promising program, praising the efforts of the student and his supervisor. He stressed that the University, under the guidance of HRH Princess Sumaya bint El Hassan, Chair of its Board of Trustees, was keen that the program succeed at the international level. He explained that 10 students were selected from among 87 students who applied for the program, with graduates representing all of Jordan’s universities. The successful candidates received full scholarships from the University, which treated them like faculty members in terms of covering the costs of their participation in scientific conferences. He pointed out that they were the top students in their specializations at the undergraduate and Master's levels, and that most of them have thus far published the equivalent of seven papers in Scopus-accredited journals.
Prof. Al-Refai explained that PSUT also provided instructors to teach and supervise the program from major international universities, such as the Indian Institute of Technology and the University of Michigan, USA.
The Dean of the King Abdullah I School of Postgraduate Studies and Scientific Research, Prof. Mohammad Sababha, stated that PSUT’s PhD program in Computer Science has been planned in such a way as to prepare researchers capable of joining research work as soon as they graduate, by supporting them to participate in international conferences and granting them research incentives to encourage them to publish in prestigious journals. This has led to the publication of almost one hundred research papers prior to their graduation.
The dissertation, which presented an innovative way to summarize Arabic texts, includes three subsystems: extracting relationships related to semantics of words from texts based on Arabic grammar and representing them using interconnected networks of words and concepts, re-representing these networks as mathematical vectors, depending on the importance of the extracted relationships, and using deep neural networks to create text summaries that contain the main concepts and ideas of the original text.
The defense committee comprised Dr. Adnan Shaout, Dr. Qasem Al-Radaideh, Dr. Sarah Tadmouri, and Dr. Salam Freihat.