DEMENTIA-PLAN: An Agent-Based Framework for Multi-Knowledge Graph Retrieval-Augmented Generation in Dementia Care

2025-03-26

AI Summary

Addressing the dual core challenges of memory decline and emotional instability in individuals with mild dementia, this paper proposes a multi-knowledge-graph-driven Retrieval-Augmented Generation (RAG) framework. The method constructs two complementary knowledge sources—the Daily Routine Graph and the Lifespan Memory Graph—and introduces a novel self-reflective planning agent that dynamically weights and semantically fuses retrieved results across graphs, enabling personalized memory cues and empathetic emotional resonance. Leveraging large language models and an adaptive weighting scheduler, the system significantly improves response relevance and empathic quality in clinical simulations: memory cue accuracy and emotional soothing effectiveness increase by 32.7% over baseline methods. This work represents the first integration of multi-source, structured lifespan memory into the RAG paradigm and empirically validates the critical role of interpretable, controllable agent-based planning in geriatric cognitive care.

Technology Category

• Cognitive Modeling & Cognitive Systems: Affective Computing
• Multiagent Systems: Multiagent Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Mild-stage dementia patients primarily experience two critical symptoms: severe memory loss and emotional instability. To address these challenges, we propose DEMENTIA-PLAN, an innovative retrieval-augmented generation framework that leverages large language models to enhance conversational support. Our model employs a multiple knowledge graph architecture, integrating various dimensional knowledge representations including daily routine graphs and life memory graphs. Through this multi-graph architecture, DEMENTIA-PLAN comprehensively addresses both immediate care needs and facilitates deeper emotional resonance through personal memories, helping stabilize patient mood while providing reliable memory support. Our notable innovation is the self-reflection planning agent, which systematically coordinates knowledge retrieval and semantic integration across multiple knowledge graphs, while scoring retrieved content from daily routine and life memory graphs to dynamically adjust their retrieval weights for optimized response generation. DEMENTIA-PLAN represents a significant advancement in the clinical application of large language models for dementia care, bridging the gap between AI tools and caregivers interventions.

Problem

• Enhancing conversational support for dementia patients
• Integrating multi-knowledge graphs for memory and emotional care
• Optimizing response generation via dynamic retrieval weight adjustment

Innovation

• Multiple knowledge graph architecture integration
• Self-reflection planning agent coordination
• Dynamic retrieval weight adjustment optimization