Innovative AI Solutions for Social Impact

It is often challenging for non-profit or civil society organizations to access scientific knowledge and expertise that could benefit the intended mission and desired impact. Many scientists and researchers are also eager to use their expertise to make a positive contribution to the world. “The Good Scientists” was born out of these two parallel realizations.

The Good Scientists provides a matchmaking platform to connect academic researchers with social and environmental impact projects.  The end goal is to create new and unconventional collaborations and make science actionable for social impact.   

 We believe that the young generation of scientists are highly motivated to make real, tangible impact in their careers. There just aren’t many opportunities to do so. That’s why we want to improve opportunities for these scientists to make social impact across all fields and stages of training.” Elena del Pup, founder The Good Scientists

 The goal: Improved search results 

The Good Scientists is a matchmaking platform. So, the goal is to present a list of researchers or scientists from various scientific fields and career stages that match the needs of an organization.  To get a list of search results, the platform uses regular filtering options. This does not always work accurately; it could be much better. But how? 

Deployment of Large Language Models  

To ensure that the match making platform will work more efficiently and accurately, The Good Scientists has developed a decentralized online ecosystem. This ecosystem consists of two components: a marketplace of Science4Good projects provided by civic organizations and a free automated booking system to book consultations with their network of scientists. 

As a solution, the developers of The Good Scientists presented training and fine-tuning Large Language Models using High Performance Computing to enable scalability and thus access to larger models. In doing so, they focus on reducing the response time for recommendations through efficient workload balancing and efficient use of GPU-resources. 

The use of Large Language Models enables dynamic analysis of textual data from multiple sources. In addition, it extracts the relevant information with a high degree of personalization and generates qualitative semantic understanding. The result? Better search results for organizations..  

Better search results, more impact

The use of a Large Language Model (LLM)-driven matching platform is an important step forward in making science accessible to social enterprises and non-profits. The use of High Performance Computing has enabled The Good Scientists’ platform to use technologies that have improved the accuracy of the search function. In addition, it has also been able to reduce the response time, generating high-quality results quickly.

More use cases