ADB Sandbox Opportunity: Artificial Intelligence for Predicting Resettlement Outcomes

Submit your solution to enhance remote monitoring of resettlement projects by utilizing Artificial Intelligence.

groups

Teams of
2 - 8 people

The Challenge

Implementation of ADB's resettlement programs is facing challenges because of outdated modes of data collection in resettlement surveys, which are largely paper-based and require many processing steps. These surveys play an integral role in resettlement planning, implementation, monitoring, and evaluation as part of the project's lifecycle.

Field-based evaluations also experienced serious disruptions during the COVID-19 pandemic. Due to heavy reliance on monitoring reports and site visits, project teams are often unable to efficiently identify and respond in a timely manner to issues that may result in complaints during resettlement implementation.

Moreover, there is no system to capture knowledge and lessons learned from resettlement programs, nor a systematic method to compare and analyze resettlement outcomes across projects.

ADB wants to hear your company’s proposed solution for a platform that combines digital data capture and real-time resettlement monitoring with an AI-driven model to predict positive or negative outcomes of resettlement programs in ADB projects.

This opportunity is under the ADB Digital Innovation Sandbox Program. Click here to learn more about available opportunities in the program.

Expected Deliverables

We are seeking to pilot test a digital tool to perform resettlement impact assessment as well as real-time monitoring, and post resettlement evaluation of resettlement programs in Mongolia. Combined with field data, synthetic data will be used to develop a predictive model for resettlement outcomes. This initiative also aims to reduce physical presence in the field especially in pandemic and emergency situations.

Real-time Monitoring

The solution must have the capability to perform real-time resettlement monitoring and provide real-time data availability to key stakeholders. This would allow greater transparency and faster analysis upon data entry in resettlement planning and implementation.

Predictive Model

A predictive model must be developed for resettlement outcomes based on machine learning and using a combination of real-world data and synthetic data.

Demonstrate Accuracy

It is expected that the tool will demonstrate increased accuracy as more project data is entered into the platform.