Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces Zero-Ingestion ScalableRAG, a novel approach that eliminates the need for costly knowledge ingestion while achieving superior performance in retrieval-augmented generation (RAG) tasks. By utilizing a dynamic workspace for document and value sets, the method enables efficient on-the-fly aggregative reasoning, outperforming traditional knowledge graph methods by an average of 7.36% across six datasets. Additionally, the Limited-Ingestion ScalableRAG variant incorporates a minimal vector database to enhance accuracy without significantly increasing costs.
Achieving high-quality retrieval-augmented generation without any ingestion costs could revolutionize how we approach knowledge integration in AI systems.
Recent advances in RAG aim to optimize for performance by paying high ingestion costs for knowledge ingestion: building knowledge graphs or extracting SQL tables. In this work we show that the operations that such knowledge bases allow can be replicated with zero ingestion costs (not even a vector database); in fact our solution, Zero-Ingestion ScalableRAG, handily out-performs all baselines (including knowledge graph approaches) in three out of the six corpora considered here, and only marginally missing maximum performance on the other three, with average accuracy across all six datasets 7.36% above the next most competitive baseline. It achieves this by keeping a workspace of document sets and values sets that it can write into and read from, allowing for on-the-fly aggregative reasoning in all situations where grouping is required on a primary key that is in one to one correspondence with a subset of the total document set. Capping the number of LLM calls by a constant independent of the corpus size, we also introduce Limited-Ingestion ScalableRAG, which does use a minimal vector database as well as an automated pattern discovery from a sample of documents, to further improve accuracy at scale. Our code is available at https://github.com/cohesity/ScalableRAG .