Semantic Anchoring in Agentic Memory: Leveraging Linguistic Structures for Persistent Conversational Context

2025-08-18

AI Summary

Existing retrieval-augmented generation (RAG) systems encode dialogue history into semantic vectors, offering generalization but neglecting fine-grained linguistic structures—such as syntactic dependencies, discourse relations, and coreferential links—thereby limiting large language models’ memory precision and contextual traceability in long-horizon, multi-turn dialogues. To address this, we propose Semantic-Anchored Memory (SAM), the first architecture to explicitly integrate dependency parsing, discourse relation labeling, and coreference resolution into memory representation, yielding structured memory entries that jointly optimize with RAG. Evaluated on long-horizon dialogue benchmarks, SAM improves factual recall and discourse coherence by up to 18%. Its effectiveness is rigorously validated through ablation studies, human evaluation, and detailed error analysis.

Technology Category

• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

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
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Large Language Models (LLMs) have demonstrated impressive fluency and task competence in conversational settings. However, their effectiveness in multi-session and long-term interactions is hindered by limited memory persistence. Typical retrieval-augmented generation (RAG) systems store dialogue history as dense vectors, which capture semantic similarity but neglect finer linguistic structures such as syntactic dependencies, discourse relations, and coreference links. We propose Semantic Anchoring, a hybrid agentic memory architecture that enriches vector-based storage with explicit linguistic cues to improve recall of nuanced, context-rich exchanges. Our approach combines dependency parsing, discourse relation tagging, and coreference resolution to create structured memory entries. Experiments on adapted long-term dialogue datasets show that semantic anchoring improves factual recall and discourse coherence by up to 18% over strong RAG baselines. We further conduct ablation studies, human evaluations, and error analysis to assess robustness and interpretability.

Problem

• Improving memory persistence in multi-session LLM conversations
• Capturing linguistic structures beyond semantic similarity
• Enhancing recall of nuanced context-rich dialogue exchanges

Innovation

• Hybrid memory architecture with linguistic cues
• Dependency parsing and coreference resolution integration
• Structured memory entries improving recall coherence