CAMI framework optimizes RAG pipelines with cost-aware multi-indexing

Researchers have developed CAMI (Cost-Aware Agent-Guided Multi-Indexing), a new framework designed to optimize the process of creating semantic enrichment indices for retrieval-augmented generation (RAG) pipelines. This framework addresses the computational bottleneck caused by the combinatorial nature of enrichment types and generator models. CAMI formalizes index construction as a budgeted, multi-objective portfolio selection problem, enabling efficient identification of high-recall portfolios under strict budget constraints. Evaluations show CAMI can achieve up to 9.4% higher recall@10 compared to content-only baselines and uses up to 5x less budget than random search methods, making it practical for production environments. AI

IMPACT Optimizes RAG pipelines, potentially improving search efficiency and recall in AI applications.

RANK_REASON The item is a research paper detailing a new framework for semantic retrieval. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Litmaps
• Retrieval-Augmented Generation
• ScienceCast
• scite Smart Citations

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →