Direct Retrieval-augmented Optimization: Synergizing Knowledge Selection and Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915275246403584 |
|---|---|
| author | Shi, Zhengliang Yan, Lingyong Sun, Weiwei Feng, Yue Ren, Pengjie Ma, Xinyu Wang, Shuaiqiang Yin, Dawei de Rijke, Maarten Ren, Zhaochun |
| author_facet | Shi, Zhengliang Yan, Lingyong Sun, Weiwei Feng, Yue Ren, Pengjie Ma, Xinyu Wang, Shuaiqiang Yin, Dawei de Rijke, Maarten Ren, Zhaochun |
| contents | Retrieval-augmented generation (RAG) integrates large language models ( LLM s) with retrievers to access external knowledge, improving the factuality of LLM generation in knowledge-grounded tasks. To optimize the RAG performance, most previous work independently fine-tunes the retriever to adapt to frozen LLM s or trains the LLMs to use documents retrieved by off-the-shelf retrievers, lacking end-to-end training supervision. Recent work addresses this limitation by jointly training these two components but relies on overly simplifying assumptions of document independence, which has been criticized for being far from real-world scenarios. Thus, effectively optimizing the overall RAG performance remains a critical challenge.
We propose a direct retrieval-augmented optimization framework, named DRO, that enables end-to-end training of two key components: (i) a generative knowledge selection model and (ii) an LLM generator. DRO alternates between two phases: (i) document permutation estimation and (ii) re-weighted maximization, progressively improving RAG components through a variational approach. In the estimation step, we treat document permutation as a latent variable and directly estimate its distribution from the selection model by applying an importance sampling strategy. In the maximization step, we calibrate the optimization expectation using importance weights and jointly train the selection model and LLM generator. Our theoretical analysis reveals that DRO is analogous to policy-gradient methods in reinforcement learning. Extensive experiments conducted on five datasets illustrate that DRO outperforms the best baseline with 5%-15% improvements in EM and F1. We also provide in-depth experiments to qualitatively analyze the stability, convergence, and variance of DRO. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_03075 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Direct Retrieval-augmented Optimization: Synergizing Knowledge Selection and Language Models Shi, Zhengliang Yan, Lingyong Sun, Weiwei Feng, Yue Ren, Pengjie Ma, Xinyu Wang, Shuaiqiang Yin, Dawei de Rijke, Maarten Ren, Zhaochun Information Retrieval Retrieval-augmented generation (RAG) integrates large language models ( LLM s) with retrievers to access external knowledge, improving the factuality of LLM generation in knowledge-grounded tasks. To optimize the RAG performance, most previous work independently fine-tunes the retriever to adapt to frozen LLM s or trains the LLMs to use documents retrieved by off-the-shelf retrievers, lacking end-to-end training supervision. Recent work addresses this limitation by jointly training these two components but relies on overly simplifying assumptions of document independence, which has been criticized for being far from real-world scenarios. Thus, effectively optimizing the overall RAG performance remains a critical challenge. We propose a direct retrieval-augmented optimization framework, named DRO, that enables end-to-end training of two key components: (i) a generative knowledge selection model and (ii) an LLM generator. DRO alternates between two phases: (i) document permutation estimation and (ii) re-weighted maximization, progressively improving RAG components through a variational approach. In the estimation step, we treat document permutation as a latent variable and directly estimate its distribution from the selection model by applying an importance sampling strategy. In the maximization step, we calibrate the optimization expectation using importance weights and jointly train the selection model and LLM generator. Our theoretical analysis reveals that DRO is analogous to policy-gradient methods in reinforcement learning. Extensive experiments conducted on five datasets illustrate that DRO outperforms the best baseline with 5%-15% improvements in EM and F1. We also provide in-depth experiments to qualitatively analyze the stability, convergence, and variance of DRO. |
| title | Direct Retrieval-augmented Optimization: Synergizing Knowledge Selection and Language Models |
| topic | Information Retrieval |
| url | https://arxiv.org/abs/2505.03075 |