GenEGM
GenEGM: Automatically analyze current global development knowledge for future policy decisions
Introduction
Our project aims to revolutionize evidence-based decision-making in global development by harnessing the power of fine-tuned large language models. Our model will generate rapid and resource-efficient evidence syntheses called evidence gap maps (EGMs), allowing policy makers to stay up-to-date with the ever-expanding body of evidence, enabling them to make the most impactful aid policy decisions without the lengthy wait for traditional evidence synsthesis tools. Furthermore, our project will prioritize equity by incorporating academic articles from the communities that funding from EGMs may impact.
Problem
When designing evidence-based policies, decision-makers must analyze a vast and rapidly growing literature base. This time and resource intensive process relies on human judgement. Accelerating this process would promote timely evidence-based decision making in global development.
Opportunity
Our solution incorporates 3 AI methods: 1) Large Language Models (LLM) to understand the existing literature base 2) Classifiers to synthesize the text content from the LLM 3) Generative language models to create an evidence gap map to suggest knowledge gaps and impact areas in global development.