Answering Clean Tech Questions with Large Language Models

The success of the “net zero transition” relies on the acceleration of the clean technology development to increase renewable energy capacity and low-emission solutions, but also to improve energy efficiency and enable carbon capture. Tracking such technologies and their mineral requirements is becoming increasingly important, but has traditionally required expert knowledge. 

In this paper, we propose a framework using Large Language Models and question-answering tasks to monitor the novelty within the clean tech industry, but also the minerals on which these technologies rely. It demonstrates the benefits of using artificial intelligence, and more specifically NLP techniques, to reconstruct expert knowledge and track rapidly changing markets.

The past few years have seen a growing number of promising net zero commitments from both governments and companies. But there may be some bumps on the road toward a smooth transition to net zero. Indeed, turning words into action has raised questions about the capacity of our economies to make such structural changes, which in turn depend on non-infinite supplies of capital, labour and technological progress. The development of clean energy is a prerequisite for achieving such ambitious goals by reducing greenhouse gas emissions (IPCC, 2023), alongside low emission fuels, carbon capture or zero-emission tech nologies such as nuclear fusion, as John Kerry proposed during COP28.

You can now read the full whitepaper at the link below