Developing AI-Informed Vulnerability Index to Target Aid

 Eligibility and Distribution for

the Food Security Program in Zambia

Image by SpaceX
Image by SpaceX

Image by SpaceX
Image by SpaceX

1/1

In responding to surging food insecurity in the wake of COVID-19, the Zambian government plans to triple the food security pack to the vulnerable population. However, identifying areas of extreme poverty has high exclusion and inclusion errors, given the sparsity of the ground-truth surveys. To address this issue, this study aims to generate robust wealth estimates at a granular level, combining satellite imagery, social media, and spatial analysis. Leveraging an AI/ML-informed vulnerability index, the Zambian governments can prioritize their aid to the neediest communities. This index can be updated as new data becomes available to decisionmakers in near-real-time.

..................................................................................................................................................................................

NUMBER OF RECIPIENTS FOR THE FOOD SECURITY PACK

chart.png

In response to increasing food insecurity, exacerbated by COVID-19, Zambia is tripling the reach of its Food Security Program.        

Image by Atlas Green

Research

.................................................

 

Dr. Jung's latest research project aims to develop a multimodal approach that combines daytime satellite imagery and social media to estimate the wealth and livelihood of regions and shape aid distribution.



              PARTNERS              ....................................................

Microsoft logo.png

Recipient of Microsoft's
AI for Humanitarian Action Project Initiatives'  $25,000 Azure Grant. 

              PARTNERS              ....................................................

              PARTNERS              ....................................................

Logo - IPA-Africa.png
Research .png

Recipient of the Research Council's Collaborative Multidisciplinary Award

Close up Portrait
Close up Portrait

Close up Portrait
Close up Portrait

1/1

Each year, billions of dollars are allocated internationally to alleviate poverty. Our work can contribute globally by offering novel sources of data for Zambia and neighboring countries to expand their social protection programs. After this project validates its performance in Zambia, we can apply the model to the Republic of the Congo, a neighboring country without the Demographic and Health Surveys (DHS) wealth data and in an even more challenging data environment.

            RESEARCH TEAM       .................................................................​​

The team consists of computer/data scientists

and social scientsts:

  • Woojin Jung - Rutgers University

  • Jordan Steiner - Rutgers University

  • Alisha Cupid - Rutgers University

  • Melissa Yu - Rutgers University

  • Amanda Starcev - Rutgers University

           3 STAGES        ................................................

  1. Assess the current machine learning model

  2. Develop new algorithms with a wider application

  3. Share this new field-tested and evaluated approach with policymakers

Research Duration
.................................................
 
July 2021 - June 2022

The study is slated to conclude in June 2022 with the dissemination of findings to policymakers.

            RESEARCH TEAM       .................................................................​​

  • Matthew Purri - Rutgers University

  • Fangzheng Wu - Rutgers University

  • Lily Lin - Microsoft

  • Dimitri Metaxa - Rutgers University

  • Tawfiq Ammar - Rutgers University

  • Dimitrios Ntarlagianni - Rutgers University

  • Owen Siyoto - Zambia Statistics Agency