Improving Multi-turn Dialogue Consistency with Self-Recall Thinking

• URL: http://arxiv.org/abs/2605.15102v1
• Date: Thu, 14 May 2026 17:20:14 GMT
• Title: Improving Multi-turn Dialogue Consistency with Self-Recall Thinking
• Abstract summary: Large language model (LLM) based multi-turn dialogue systems often struggle to track dependencies across non-adjacent turns.<n>We propose Self-Recall Thinking (SRT), a framework designed to address long-range contextual dependency and sparse informative signals.<n> Experiments on multiple datasets demonstrate that SRT improves F1 score by 4.7% and reduces end-to-end latency by 14.7% over prior methods.
• Score: 13.308970592702602
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language model (LLM) based multi-turn dialogue systems often struggle to track dependencies across non-adjacent turns, undermining both consistency and scalability. As conversations lengthen, essential information becomes sparse and is buried in irrelevant context, while processing the entire dialogue history incurs severe efficiency bottlenecks. Existing solutions either rely on high latency external memory or lose fine-grained details through iterative summarization. In this paper, we propose Self-Recall Thinking (SRT), a framework designed to address long-range contextual dependency and sparse informative signals in multi-turn dialogue. SRT identifies helpful historical turns and uses them to generate contextually appropriate responses, enabling the model to selectively recall and reason over context during inference. This process yields an endogenous reasoning process that integrates interpretable recall steps without external modules. SRT incorporates: (1) Dependency Construction: Generating and converting it into self-recall chains; (2)Capability Initialization: Training to enable reasoning chains with recall tokens capability; (3)Reasoning Improvement: Refining accuracy via verifiable rewards to optimize recall and reasoning for correct answers. Experiments on multiple datasets demonstrate that SRT improves F1 score by 4.7% and reduces end-to-end latency by 14.7% over prior methods, achieving a balance between reasoning latency and accuracy, and outperforming state-of-the-art baselines.

Related papers

• RetroThinker: Enabling Retrospective Thinking in Speech LLMs [42.95471264630473]
  Speech large language models (SpeechLLMs) lag behind text-only LLMs on complex reasoning tasks.<n>We introduce RetroThinker, a multi-stage post-training framework that equips the Moshi model to self-verify and forward-correct CoT steps during inference.<n>RetroThinker significantly improves the accuracy-latency trade-off over non-retrospective baselines, achieving an 11% absolute accuracy gain at a comparable latency.
  arXiv  Detail & Related papers  (2026-09-10T17:41:53Z)
• MINTEval: Evaluating Memory under Multi-Target Interference in Long-Horizon Agent Systems [69.06764269022925]
  We study how current memory-augmented agents perform in realistic, interference-heavy, long-horizon settings.<n> MINTEval is a benchmark featuring long, highly interconnected contexts with frequently updated information.<n>Overall, MINTEval has 15.6k question-answering pairs over long-horizon contexts averaging 138.8k tokens and extending to 1.8M tokens per instance.
  arXiv  Detail & Related papers  (2026-05-18T15:43:35Z)
• Back to Basics: Let Conversational Agents Remember with Just Retrieval and Generation [36.056012318684175]
  Existing conversational memory systems rely on complex hierarchical summarization or reinforcement learning to manage long-term dialogue history.<n>We identify two key phenomena: textitDecisive Evidence Sparsity and textitDual-Level Redundancy<n>Motivated by these insights, we propose method, a minimalist framework that brings conversational memory back to basics.
  arXiv  Detail & Related papers  (2026-04-13T15:38:43Z)
• Self-Correcting RAG: Enhancing Faithfulness via MMKP Context Selection and NLI-Guided MCTS [5.711356514732554]
  Self-Correcting RAG reformulates retrieval and generation as constrained optimization and path planning.<n>On the input side, we formalize context selection as a multi-dimensional multiple-choice knapsack problem (MMKP)<n>On the output side, we introduce a natural language inference (NLI)-guided Monte Carlo Tree Search (MCTS) mechanism.
  arXiv  Detail & Related papers  (2026-04-12T17:14:36Z)
• Balancing Faithfulness and Performance in Reasoning via Multi-Listener Soft Execution [79.98699884805636]
  Reasoning Execution by Multiple Listeners (REMUL) is a multi-party reinforcement learning approach.<n>REMUL builds on the hypothesis that reasoning traces which other parties can follow will be more faithful.<n>Speakers are rewarded for producing reasoning that is clear to listeners.
  arXiv  Detail & Related papers  (2026-02-18T02:55:55Z)
• Structured Reasoning for Large Language Models [59.215789462977206]
  We propose Structured Reasoning (SCR), a framework that decouples reasoning trajectories into explicit, evaluable, and trainable components.<n>SCR substantially improves reasoning efficiency and self-verification.<n>Compared with existing reasoning paradigms, it reduces output token length by up to 50%.
  arXiv  Detail & Related papers  (2026-01-12T04:04:01Z)
• Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents [80.33280979339123]
  We introduce Memory-T1, a framework that learns a time-aware memory selection policy using reinforcement learning (RL)<n>On the Time-Dialog benchmark, Memory-T1 boosts a 7B model to an overall score of 67.0%, establishing a new state-of-the-art performance for open-source models.
  arXiv  Detail & Related papers  (2025-12-23T06:37:29Z)
• Rhea: Role-aware Heuristic Episodic Attention for Conversational LLMs [36.91809943381492]
  Large Language Models (LLMs) have achieved remarkable performance on single-turn tasks, yet their effectiveness deteriorates in multi-turn conversations.<n>We propose Rhea, a novel framework that decouples conversation history into two functionally independent memory modules.<n>Experiments show that Rhea mitigates performance decay and improves overall accuracy by 1.04 points on a 10-point scale.
  arXiv  Detail & Related papers  (2025-12-07T14:50:03Z)
• Thinking Forward and Backward: Multi-Objective Reinforcement Learning for Retrieval-Augmented Reasoning [137.33138614095435]
  Retrieval-augmented generation (RAG) has proven to be effective in mitigating hallucinations in large language models.<n>Recent efforts have incorporated search-based interactions into RAG, enabling iterative reasoning with real-time retrieval.<n>We propose Bi-RAR, a novel retrieval-augmented reasoning framework that evaluates each intermediate step jointly in both forward and backward directions.
  arXiv  Detail & Related papers  (2025-11-12T08:29:39Z)
• D-SMART: Enhancing LLM Dialogue Consistency via Dynamic Structured Memory And Reasoning Tree [22.420810089099614]
  Large Language Models (LLMs) often exhibit factual inconsistencies and logical decay in extended, multi-turn dialogues.<n>We propose D--101, a model-agnostic framework designed to maintain multi-turn dialogue consistency.<n>We introduce new NLI-based metrics to better measure multiturn dialogue consistency.
  arXiv  Detail & Related papers  (2025-10-15T09:53:11Z)
• In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents [70.12342024019044]
  Large Language Models (LLMs) have made significant progress in open-ended dialogue, yet their inability to retain and retrieve relevant information limits their effectiveness.<n>We propose Reflective Memory Management (RMM), a novel mechanism for long-term dialogue agents, integrating forward- and backward-looking reflections.<n>RMM shows more than 10% accuracy improvement over the baseline without memory management on the LongMemEval dataset.
  arXiv  Detail & Related papers  (2025-03-11T04:15:52Z)
• PICK: Polished & Informed Candidate Scoring for Knowledge-Grounded Dialogue Systems [59.1250765143521]
  Current knowledge-grounded dialogue systems often fail to align the generated responses with human-preferred qualities. We propose Polished & Informed Candidate Scoring (PICK), a generation re-scoring framework. We demonstrate the effectiveness of PICK in generating responses that are more faithful while keeping them relevant to the dialogue history.
  arXiv  Detail & Related papers  (2023-09-19T08:27:09Z)

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.