Gespeichert in:
| Hauptverfasser: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2508.19282 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911725602734080 |
|---|---|
| author | Cui, Ziqiang Weng, Yunpeng Tang, Xing Liu, Peiyang Li, Shiwei He, Bowei Chen, Jiamin Zhang, Yansen He, Xiuqiang Ma, Chen |
| author_facet | Cui, Ziqiang Weng, Yunpeng Tang, Xing Liu, Peiyang Li, Shiwei He, Bowei Chen, Jiamin Zhang, Yansen He, Xiuqiang Ma, Chen |
| contents | Retrieval-Augmented Generation (RAG) has emerged as a promising paradigm for improving the timeliness of knowledge updates and the factual accuracy of large language models. However, incorporating a large volume of retrieved documents significantly increases input length, leading to prohibitive computational costs. Existing compression approaches often compromise task performance, primarily due to their reliance on predefined heuristics. These heuristics fail to ensure that the compressed context is conducive to the generation tasks. To address these limitations, we propose CORE-RAG, a novel framework for context compression in RAG systems. CORE eliminates reliance on proxy heuristics through a performance-driven learning framework, which directy utilizes task performance as a feedback signal to iteratively refine the compressor policy. Prior to this optimization process, we incorporate a knowledge distillation phase to initialize the compressor with a robust policy. Extensive experiments demonstrate the superiority of our approach. At a high compression ratio of 3%, CORE not only avoids performance degradation but also improves the average Exact Match (EM) score by 3.3 points compared to using full documents. Our code is available at https://github.com/ziqiangcui/CORE-RAG-ICML26. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_19282 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Less Is More: Elevating RAG via Performance-Driven Context Compression Cui, Ziqiang Weng, Yunpeng Tang, Xing Liu, Peiyang Li, Shiwei He, Bowei Chen, Jiamin Zhang, Yansen He, Xiuqiang Ma, Chen Computation and Language Artificial Intelligence Retrieval-Augmented Generation (RAG) has emerged as a promising paradigm for improving the timeliness of knowledge updates and the factual accuracy of large language models. However, incorporating a large volume of retrieved documents significantly increases input length, leading to prohibitive computational costs. Existing compression approaches often compromise task performance, primarily due to their reliance on predefined heuristics. These heuristics fail to ensure that the compressed context is conducive to the generation tasks. To address these limitations, we propose CORE-RAG, a novel framework for context compression in RAG systems. CORE eliminates reliance on proxy heuristics through a performance-driven learning framework, which directy utilizes task performance as a feedback signal to iteratively refine the compressor policy. Prior to this optimization process, we incorporate a knowledge distillation phase to initialize the compressor with a robust policy. Extensive experiments demonstrate the superiority of our approach. At a high compression ratio of 3%, CORE not only avoids performance degradation but also improves the average Exact Match (EM) score by 3.3 points compared to using full documents. Our code is available at https://github.com/ziqiangcui/CORE-RAG-ICML26. |
| title | Less Is More: Elevating RAG via Performance-Driven Context Compression |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2508.19282 |