CDS4RAG: Cyclic Dual-Sequential Hyperparameter Optimization for RAG
Pengzhou Chen, Tao Chen · 2026
Read CDS4RAG: Cyclic Dual-Sequential Hyperparameter Optimization for RAG by Pengzhou Chen, Tao Chen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-08333.txt.
Description
Pengzhou Chen and Tao Chen tune retrieval-augmented generation systems as one coupled problem. A RAG pipeline carries many settings. They split between its retriever and its generator. The two sides interact as a pair. Tuning either side alone leaves gains on the table. Testing full setups is slow and costly too. A joint method has to be smart about where it spends its time.
Their CDS4RAG framework gives the problem a cyclic dual-sequential form. Retriever settings and generator settings are tuned in turns. Each cycle inherits seeds from the last. Fine-grained budget rules spend tests where they help most. The framework is algorithm-agnostic. Other general optimizers can plug in. The authors report faster convergence than black-box approaches.
Catalog record
| CREATOR | Pengzhou Chen, Tao Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.08333