October 8, 2025Open Access

Retrieval-Augmented Generation: A Comprehensive Survey of Architectures, Enhancements, and Robustness Frontiers

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This survey highlights advancements in retrieval-augmented generation, addressing retrieval quality and robustness challenges.

Retrieval-augmented generation improves large language models by using external evidence, tackling factual inconsistencies and domain rigidity.
The survey categorizes architectures into retriever-centric, generator-centric, and hybrid models, examining their enhancements in efficiency and performance.
A comprehensive analysis of performance on multi-hop question answering reveals trade-offs among retrieval precision and generation flexibility.
Open challenges for future research include adaptive retrieval architectures and privacy-preserving mechanisms to enhance robustness.

Chhavi Sharma (2025) studied this question.

View Full Paper

Ask AI

Bookmark