Close
Email: adham.enaya.22@ucl.ac.uk
Thesis Title: Improving the Performance of Deep Gravity Models in Predicting Urban Mobility in Geographic Areas with Limited Training Data through the Use of Urban Transfer Learning Techniques
Primary Supervisor: Zhong, Chen
Second Supervisor: Michael Batty
Funding Source: realTRIPS project
Start Date: February 2023
Biography
Adham is a doctoral researcher in urban mobility analysis and modeling at the Bartlett Centre for Advanced Spatial Analysis. He holds a Bachelor's degree in Software Engineering from Al-Azhar University in Palestine and a Master's in Urban Analytics and Visualisation from Warwick University in the UK. Prior to his doctoral studies, he worked as a senior software developer and analyst at the United Nations in the Middle East, where he leveraged his expertise to support the agency's efforts in delivering essential services and empowering the refugee communities through technology solutions.
Research Summary
The research focuses on improving the performance of Gravity Models that use deep learning to generate flow probabilities in geographic areas with limited training data. The issue addressed is the low generalization capabilities of neural networks trained on spatial data when applied to different regions. The proposed solution is to utilize Urban Transfer Learning, a machine learning technique that transfers knowledge from a source city with rich data to a target city with limited data.
Research Themes
Urban Mobility, Spatial Data Science, Transfer Learning, GIS, Research Software Engineering