Dr. Fatima Abu Salem
|
Associate Professor
American University of Beirut (AUB) |
|
Dr. F. K. Abu Salem is associate professor of Computer Science at the American University of Beirut. She has an MS in pure Mathematics from AUB and a DPhil in Computing from the University of Oxford. Dr. Abu Salem’s former research area has been in Computer Algebra, with a focus on developing parallel and cache efficient algorithmic designs for problems in polynomial factorization and irreducibility testing over finite fields. Currently, she is involved in data science research for the public good. Applications that she has been tackling include but are not limited to, automatic fake news detection in the Syrian Conflict, predicting demand on the primary health care centers in Lebanon as a result of the Syrian refugee influx, and predicting refugee mobility in Turkey as a result of large-scale events, and mining for historical trends in bias discourse around various conflicts arising during the Lebanese war, using leading Lebanese newspapers archives.
|
Dr. Abu Salem is also co-leading predictive analytics in collaboration with the Medical and Public Health School at AUB, in two projects related to the prediction of dementia among elderly patients as well as the prediction of birth defects among Lebanese women. Most recently, Dr. Abu Salem is a co-awardee of a $1,000,000 grant from Google’s AI Impact challenge for the social good. In this project Dr. Abu Salem is exploring advanced probabilistic machine learning models to predict Evapotranspiration metrics that can help farmers optimize on their irrigation practices, thereby improving on water and energy consumption as well as improving crop yields.
Dr. Abu Salem is also leading a university wide attempt to incorporate machine learning into predicting and prescribing Freshman student performance outcomes, using students’ transcripts of grades at AUB and data pre-admission to AUB. Dr. Abu Salem has spoken multiple times on challenges affecting women in Computing at ARAB-WIC events (Arab Women in Computing network). She has served as secretary for the special activity group on Supercomputing for the Society of Industrial and Applied Mathematics, and she currently serves as associate editor for the Journal of Parallel and Distributed Computing.
Technical Vision Talk: Doing Data Science in Data Deserts
In certain developing countries, data abundance and accessibility to it may be too much of a luxury. In the absence of governmental support for promoting the data revolution, researchers turn to grassroots initiatives to obtain viable data. Though limited in each of the 3Vs – volume, velocity, and variety -- such data can still reveal interesting phenomena, and advanced machine learning techniques, such as imbalanced learning and a few-shot learning, can improve predictive power, when data is not-so-big. In the first part of my talk, I report on a series of several related works associated with the Syrian conflict, with the help of data obtained from the Violations Documentation Center (VDC). For example, I present results on fake news detection, predicting primary health care demand by Syrian refugees in Lebanon, and understanding some notions of Syrian refugee mobility in Turkey, all seen as instigated by "peaks" in the Syrian war, revealed through the VDC. In the second part of my talk, I present a brief overview of in-progress projects with a social impact, in application to smart irrigation, predicting birth defects in Lebanon using air pollution data, and quantifying anti-refugee bias across Lebanese news corpora.
Meet-the-Speaker:
Fatima will also participate in a Meet the Speaker session, where you will be able to have a deeper dive conversation about topics covered on the main stage. Accessible to WiDS Worldwide registrants.
Fatima will also participate in a Meet the Speaker session, where you will be able to have a deeper dive conversation about topics covered on the main stage. Accessible to WiDS Worldwide registrants.