Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAG

• URL: http://arxiv.org/abs/2609.05152v1
• Date: Fri, 04 Sep 2026 13:53:43 GMT
• Title: Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAG
• Abstract summary: We propose DEX-Comp, a two-stage training recipe for RAG inference.<n>DEX-Comp compresses retrieved contexts by $16times$ and accelerates inference by $4times$--$24times$.<n>On five open-domain QA benchmarks at retrieval depths, DEX-Comp compresses retrieved contexts by $16times$ and accelerates inference by $4times$--$24times$.
• Score: 24.94881830262324
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Retrieval-Augmented Generation (RAG) enhances language models with external knowledge, but the lengthy retrieved context inflates the input and degrades inference efficiency. Soft context compression encodes each document into a substantially shorter embedding sequence. However, most existing approaches are trained by distilling outputs from uncompressed RAG systems, inherently limiting their performance relative to the original model. To address this limitation, we propose DEX-Comp, a two-stage training recipe: Pure Distillation warm-starts the compression model on the uncompressed RAG's correct responses only, and Hard Exploration then runs reinforcement learning solely on queries the uncompressed RAG fails, forcing the model to explore computation patterns better suited to compressed representations. On five open-domain QA benchmarks at retrieval depths from top-5 to top-30, DEX-Comp compresses retrieved contexts by $16\times$ and accelerates inference by $4\times$--$24\times$, while achieving performance comparable to or exceeding the uncompressed RAG baseline across retrieval depths. Ablations and evaluations across diverse datasets and backbones further confirm the contribution of each stage and the generalization of our approach.

Related papers

• RAGOCR: Optical Compression of Retrieval-Augmented Text via Visual Representation [7.540266838459537]
  RAGOCR is a novel framework that compresses retrieved documents into compact visual representations conditioned on the input query.<n>We show that RAGOCR surpasses naive RAG by over 15% in accuracy while requiring only one-eighth the number of input tokens.
  arXiv  Detail & Related papers  (2026-08-01T16:59:49Z)
• CLAD: Efficient Log Anomaly Detection Directly on Compressed Representations [2.487363589269089]
  We introduce CLAD, the first deep learning framework to perform LAD directly on compressed byte streams.<n>We propose a purpose-built architecture integrating a dilated convolutional byte encoder, a hybrid Transformer--mLSTM, and four-way aggregation pooling.<n> evaluated across five datasets, CLAD achieves a state-of-the-art average F1-score of 0.9909 and outperforms the best baseline by 2.72 percentage points.
  arXiv  Detail & Related papers  (2026-04-14T17:57:01Z)
• Arbitrary Ratio Feature Compression via Next Token Prediction [52.10426317889982]
  Arbitrary Ratio Feature Compression (ARFC) framework supports any compression ratio with a single model.<n>ARC is an auto-regressive model that performs compression via next-gressive prediction.<n>MoS module refines the compressed tokens by utilizing multiple compression results.<n>ERGC is integrated into the training process to preserve semantic and structural relationships during compression.
  arXiv  Detail & Related papers  (2026-02-12T02:38:57Z)
• Proxy Compression for Language Modeling [58.904023114033954]
  proxy compression is an alternative training scheme that preserves the efficiency benefits of compressed inputs.<n>Experiments on code language modeling demonstrate that proxy compression substantially improves training efficiency.<n>As model scale increases, proxy-trained models eventually match or rival tokenizer approaches.
  arXiv  Detail & Related papers  (2026-02-04T07:36:46Z)
• Rethinking Autoregressive Models for Lossless Image Compression via Hierarchical Parallelism and Progressive Adaptation [75.58269386927076]
  Autoregressive (AR) models are often dismissed as impractical due to prohibitive computational cost.<n>This work re-thinks this paradigm, introducing a framework built on hierarchical parallelism and progressive adaptation.<n> Experiments on diverse datasets (natural, satellite, medical) validate that our method achieves new state-of-the-art compression.
  arXiv  Detail & Related papers  (2025-11-14T06:27:58Z)
• Simple Context Compression: Mean-Pooling and Multi-Ratio Training [12.049015994907629]
  We develop a lightweight and simple mean-pooling approach that consistently outperforms the widely used compression-tokens architecture.<n>We conduct extensive experiments across in-domain and out-of-domain QA datasets, as well as across model families, scales, and compression ratios.<n>Overall, our simple mean-pooling approach achieves the strongest performance, with a relatively small drop when training for multiple compression ratios.
  arXiv  Detail & Related papers  (2025-10-23T17:57:23Z)
• CORE-RAG: Lossless Compression for Retrieval-Augmented LLMs via Reinforcement Learning [22.93037884068796]
  Retrieval-Augmented Generation (RAG) has emerged as a promising approach to enhance the timeliness of knowledge updates and the factual accuracy of responses in large language models.<n>Existing approaches to document compression tailored for RAG often degrade task performance.<n>We propose CORE, a novel method for lossless context compression in RAG.
  arXiv  Detail & Related papers  (2025-08-24T12:21:50Z)
• Enhancing RAG Efficiency with Adaptive Context Compression [24.94881830262324]
  Retrieval-augmented generation incurs significant inference costs due to lengthy retrieved contexts.<n>Existing methods apply fixed compression rates, over-compressing simple queries or under-compressing complex ones.<n>We propose Adaptive Context Compression for RAG (ACC-RAG), a framework that dynamically adjusts compression rates based on input complexity.
  arXiv  Detail & Related papers  (2025-07-24T13:46:51Z)
• BRIEF: Bridging Retrieval and Inference for Multi-hop Reasoning via Compression [91.23933111083389]
  Retrieval-augmented generation (RAG) can supplement large language models (LLMs) by integrating external knowledge.<n>This paper presents BRIEF, a lightweight approach that performs query-aware multi-hop reasoning.<n>Based on our synthetic data built entirely by open-source models, BRIEF generates more concise summaries.
  arXiv  Detail & Related papers  (2024-10-20T04:24:16Z)
• xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token [108.7069350303884]
  xRAG is an innovative context compression method tailored for retrieval-augmented generation.<n>xRAG seamlessly integrates document embeddings into the language model representation space.<n> Experimental results demonstrate that xRAG achieves an average improvement of over 10% across six knowledge-intensive tasks.
  arXiv  Detail & Related papers  (2024-05-22T16:15:17Z)
• Learning Accurate Performance Predictors for Ultrafast Automated Model Compression [86.22294249097203]
  We propose an ultrafast automated model compression framework called SeerNet for flexible network deployment. Our method achieves competitive accuracy-complexity trade-offs with significant reduction of the search cost.
  arXiv  Detail & Related papers  (2023-04-13T10:52:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.