Large language model retrieval augmented generation for energy demand consultations and beyond

We’re developing a database-driven solution that combines existing large language model (LLM) architectures with retrieval augmented generation (RAG).

In practice, this means taking EDRC and CREDS research outputs and converting them into a searchable database, allowing an LLM to access and draw on our research directly and accurately when generating responses.

We’re currently assembling the database and developing LLM interface.

• Database: Assembly of a data base of key EDRC and CREDS research outputs.
• Manual use: Researchers should be able to query the system directly to learn what our database says about specific issues.
• Automated use: The system should be able to analyse government consultations and highlight the most relevant EDRC/CREDS materials for each question.

Collect existing knowledge and papers into an embeddings database

Develop processes to update manually and automatically as new material becomes available.

Integrate an LLM with the database to identify and retrieve relevant information such that up-to-date information can be utilised for real time AI analysis and application to new consultations.

So far, we have achieved an initial set of EDRC and CREDS research outputs as a database, and an AI-enhanced search with summary function. This is a key step in the process of developing the ability analyse data in order to respond to consultations.