mlsanityData to 14 Oct 2026

AbstRAG: Learning to Abstract for Retrieval Problems

arXiv 2606.09459 ↗First posted 8 Jun 2026cs.CLcited by 0

On the map: Retrieval-augmented generation → RAG pipeline design and evaluation

Abstract (arXiv)

Retrieval-augmented generation often fails when the query, the document evidence, and the user's intent are expressed at different levels of abstraction. A query may ask about a class, a relation, or an event, while the document only states specific instances, indirect framings, or scoped formulations. We define this mismatch as an abstraction gap: the minimal set of typed assumptions required to align query intent with the available evidence. To close this gap, we introduce AbstRAG, which treats abstraction as an explicit retrieval object. AbstRAG decomposes the query--evidence gap into expression, conceptual, intent--evidence, and event-type components, and scores relevance by combining match quality, a query-independent utility prior, and the cost of the required bridges. Its central mechanism is reflective refinement: a critic diagnoses retrieval failures, localizes the failed abstraction operator, proposes a minimal stage-specific patch, and accepts the patch only under sufficiency and compression controls. Across three within-document retrieval benchmarks against seven baselines, AbstRAG outperforms on nDCG@10 in 18 of 21 paired-bootstrap contrasts and improves generation accuracy by 1.9%, 5.2%, and 4.0% across the three benchmarks; ablations confirm that reflective refinement drives most of the retrieval gain and the compression control alone reduces over-expansion false positives from 73.7% to 0% on a stress slice.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in RAG pipeline design and evaluation, one of 15 directions in Retrieval-augmented generation. Builds on 21 of its 53 references that are in the map. First posted on arXiv 8 Jun 2026.

Builds on

21 of 53 references.

References · 21

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  3. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  4. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  5. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,453
  6. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,375
  7. 2020ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERTcited by 1,178
  8. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questionscited by 923
  9. 2023C-Pack: Packed Resources For General Chinese Embeddingscited by 578
  10. 2020Fact or Fiction: Verifying Scientific Claimscited by 554
  11. 2024Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inferencecited by 554
  12. 2022Precise Zero-Shot Dense Retrieval without Relevance Labelscited by 517
9 more
  1. 2021A Dataset of Information-Seeking Questions and Answers Anchored in Research Paperscited by 486
  2. 2024Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexitycited by 451
  3. 2023Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergycited by 346
  4. 2020HoVer: A Dataset for Many-Hop Fact Extraction And Claim Verificationcited by 201
  5. 2023AVeriTeC: A Dataset for Real-world Claim Verification with Evidence from the Webcited by 144
  6. 2023Complex Claim Verification with Evidence Retrieved in the Wildcited by 110
  7. 2021AmbiFC: Fact-Checking Ambiguous Claims with Evidencecited by 38
  8. 2024Verification and Refinement of Natural Language Explanations through LLM-Symbolic Theorem Provingcited by 35
  9. 2024Controlled LLM-based Reasoning for Clinical Trial Retrievalcited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Rethinking Retrieval-Augmentation as Synthesis: A Query-Aware Context Merging Approach
  2. 2026Training the Knowledge Base through Evidence Distillation and Write-Back Enrichment
  3. 2025ArgRAG: Explainable Retrieval Augmented Generation using Quantitative Bipolar Argumentation
  4. 2026Utilizing Metadata for Better Retrieval-Augmented Generation
  5. 2026ASTRA-QA: A Benchmark for Abstract Question Answering over Documents
  6. 2025Test-time Corpus Feedback: From Retrieval to RAG
  7. 2025RAGSmith: A Framework for Finding the Optimal Composition of Retrieval-Augmented Generation Methods Across Datasets
  8. 2024Evaluating Retrieval Quality in Retrieval-Augmented Generation