Post Persona Alignment for Multi-Session Dialogue Generation

2025-06-13

AI Summary

Multi-turn role-playing dialogue faces the challenge of simultaneously maintaining long-term persona consistency and response diversity. This paper proposes a two-stage post-hoc persona alignment framework: first generating a context-aware generic response, then using it as a query to retrieve relevant persona traits from a semantic persona memory bank and inject them precisely into the response. This paradigm decouples response generation from persona modeling for the first time, avoiding rigidity and bias induced by conventional pre-generation persona injection. The method integrates large language model (LLM) inference, efficient semantic retrieval, and lightweight response refinement, enabling response-driven memory retrieval. Evaluated on multi-turn LLM dialogue benchmarks, our approach achieves significant improvements over state-of-the-art methods in persona relevance (+12.3%), consistency (+9.7%), and diversity (+8.1%).

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Multi-session persona-based dialogue generation presents challenges in maintaining long-term consistency and generating diverse, personalized responses. While large language models (LLMs) excel in single-session dialogues, they struggle to preserve persona fidelity and conversational coherence across extended interactions. Existing methods typically retrieve persona information before response generation, which can constrain diversity and result in generic outputs. We propose Post Persona Alignment (PPA), a novel two-stage framework that reverses this process. PPA first generates a general response based solely on dialogue context, then retrieves relevant persona memories using the response as a query, and finally refines the response to align with the speaker's persona. This post-hoc alignment strategy promotes naturalness and diversity while preserving consistency and personalization. Experiments on multi-session LLM-generated dialogue data demonstrate that PPA significantly outperforms prior approaches in consistency, diversity, and persona relevance, offering a more flexible and effective paradigm for long-term personalized dialogue generation.

Problem

• Maintaining long-term consistency in multi-session persona dialogues
• Preserving persona fidelity across extended conversational interactions
• Balancing response diversity and personalization in dialogue generation

Innovation

• Two-stage Post Persona Alignment framework
• Generates response before retrieving persona
• Refines response for persona consistency