mlsanityData to 14 Oct 2026

Co-Evolving Graph and Text Memory for Training-Free Multi-Hop Question Answering

arXiv 2607.23278 ↗First posted 25 Jul 2026cs.CL cs.MAcited by 0

On the map: LLM reasoning → Knowledge- and tool-augmented reasoning

Abstract (arXiv)

Multi-hop question answering requires coordinating relational and textual evidence across reasoning steps, a combination neither a text corpus nor a knowledge graph can supply alone. Prior work often emphasizes only part of this loop: graph-augmented RAG retrieves from a pre-built or query-updated graph, KGQA systems search within topic-centered subgraphs, and memory-augmented agents maintain evolving memories without continuously reconciling graph memory with textual context. We propose Co-E, a training-free system built around synchronized bidirectional graph-text working memory. A synchronization cycle consolidates textual memory, extracts relational triples into graph memory, and injects graph facts back into the generation context. Because both memories are maintained, they shape subsequent retrieval and generation. Evaluated on six multi-hop QA benchmarks, Co-E improves over comparable training-free open-backbone baselines and is competitive with larger or trained systems.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Knowledge- and tool-augmented reasoning, one of 17 directions in LLM reasoning. Builds on 27 of its 64 references that are in the map. First posted on arXiv 25 Jul 2026.

Builds on

27 of 64 references.

Closest ancestors · 6

Same direction, newest first

  1. 2026HyperGraphPro: Progress-Aware Reinforcement Learning for Structure-Guided Hypergraph RAG25 Jan
  2. 2025Multi-hop Reasoning via Early Knowledge Alignment23 Dec
  3. 2025iQUEST: An Iterative Question-Guided Framework for Knowledge Base Question Answering2 Jun
  4. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning12 Mar
  5. 2023Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph15 Jul
  6. 2022Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions20 Dec

Other references · 21

  1. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  2. 2020Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Stepscited by 1,282
  3. 2025Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Modelscited by 1,261
  4. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,075
  5. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,063
  6. 2018The Web as a Knowledge-base for Answering Complex Questionscited by 410
  7. 2024Retrieval-Augmented Generation for AI-Generated Content: A Surveycited by 354
  8. 2025Agentic Retrieval-Augmented Generation: A Survey on Agentic RAGcited by 249
  9. 2025From RAG to Memory: Non-Parametric Continual Learning for Large Language Modelscited by 231
  10. 2024Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solverscited by 138
  11. 2026Search-P1: Path-Centric Reward Shaping for Stable and Efficient Agentic RAG Trainingcited by 18
  12. 2025ComposeRAG: A Modular and Composable RAG for Corpus-Grounded Multi-Hop Question Answeringcited by 11
9 more
  1. 2026SE-Search: Self-Evolving Search Agent via Memory and Dense Rewardcited by 8
  2. 2026Meta-Reinforcement Learning with Self-Reflection for Agentic Searchcited by 8
  3. 2026GraphWalker: Agentic Knowledge Graph Question Answering via Synthetic Trajectory Curriculumcited by 6
  4. 2025KERAG: Knowledge-Enhanced Retrieval-Augmented Generation for Advanced Question Answeringcited by 5
  5. 2026MemSearch-o1: Empowering Large Language Models with Reasoning-Aligned Memory Growth in Agentic Searchcited by 5
  6. 2025SUBQRAG: Sub-Question Driven Dynamic Graph RAGcited by 3
  7. 2025KBQA-R1: Reinforcing Large Language Models for Knowledge Base Question Answeringcited by 3
  8. 2026Use Graph When It Needs: Efficiently and Adaptively Integrating Retrieval-Augmented Generation with Graphscited by 3
  9. 2025PPoGA: Predictive Plan-on-Graph with Action for Knowledge Graph Question Answeringcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Omne-R1: Learning to Reason with Memory for Multi-hop Question Answering
  2. 2026LivingRAG: Augmenting Graph RAG with Experience
  3. 2026GRASP: Graph Agentic Search over Propositions for Multi-hop Question Answering
  4. 2026BridgeRAG: Training-Free Bridge-Conditioned Retrieval for Multi-Hop Question Answering
  5. 2026LiteRAG: Cost-Efficient Graph-Based Retrieval-Augmented Generation
  6. 2026A2RAG: Adaptive Agentic Graph Retrieval for Cost-Aware and Reliable Reasoning
  7. 2025StepChain GraphRAG: Reasoning Over Knowledge Graphs for Multi-Hop Question Answering
  8. 2025Memory-Aware and Uncertainty-Guided Retrieval for Multi-Hop Question Answering