Two members of the Information Technology University (ITU) faculty recently won a competitive scholar award from Google. Dr. Mohsen Ali, an assistant professor or computer sciences and Dr. Izza Aftab, an assistant professor of economics, worked together to create a joint paper named “Is Economics from Afar Domain Generalizable?” The two professors are the only ones to have won this award from Pakistan as of 2021.
The Research Scholar Programme is a recent endeavor initiated by Google in March 2020. It aims to support early-career faculty who are committed to impactful research in fields relevant to Google. The programme will help develop new collaborations and encourage long-term international relationships within the academic community. There were participants from 50 different universities from 15 different countries. Though they were majorly universities from Europe and America, there were few from Asia, including ITU.
Dr. Mohsen Ali’s interest in this project lies in the solutions to both theoretical and practical issues entailing Computer Vision, Artificial Intelligence and Machine Learning. He is specifically interested in those problems relating to image co-segmentation, remote sensing, medical imaging, and affective computing. Dr. Izza Aftab has academic interests leaning towards the theory and modeling of firm-level innovations within still-developing countries the Economics behind climate change, and the role of Big Data in Public Policy.
Their joint paper falls under the categories of machine learning, one of the categories within which Google was interested in. Other categories of computer science study which were seen in the competition were human-computer interaction, health research, systems, and many more.
Their project involves the design of an interpretable machine learning algorithm with when combined with satellite imagery and geo-spatial datasets, can help predict economic and development indicators within specific regions.
The Project is multidisciplinary in nature; amalgamating aspects of machine learning, economics & remote sensing to development and policy. We use satellite imagery to detect slums across major cities in Pakistan. Then we go there and collect data to estimate poverty, deprivation and level of development. This helps identify, to the most micro level possible, where the next public intervention in health and education needs to be! Moreover, it gives a pretty good idea where to send transfer payments to!
Our project can guide businesses (small and medium) as well. The applications are immense and multifaceted!
Our work potentially revolutionises how data is collected and analysed in Pakistan. Moreover, and very very pertinently, it will help us & others work through important questions in the field machine learning, such as interpretability and domain generalization.
This achievement is a great accomplishment not only for the two faculty members, but ITU, Pakistan, and Google as well. The economic and technological future can be further discovered with technology that helps businesses and governments find areas in poverty and take requisite action. What follows this innovative paper and scholar programme is yet to be discovered.
Reported By: Ehtisham