AURA: Agentic Multi-Source RAG

Content

Retrieval Augmented Generation (RAG) has become a common tool for combining the abilities of large language models with one or several specific datasets. This PG focuses on questions related to the efficient usage of mutliple sources within an agentic RAG. Possible sub-topics that the PG may look at:

• RAG based on multiple sources with an intelligent way to select single sources for querying
• Agentic RAG that can trigger searches
• An approach to generate evaluation data based on shallow knowledge graphs

Note that these sub-topics might change before over time depending on the state of the art and the progress of the PG itself.

Contact