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

TL;DR

This paper introduces Semantic Anchoring, a hybrid memory system for LLMs that combines vector embeddings with linguistic structures to enhance long-term conversational recall and coherence.

Contribution

It presents a novel hybrid memory architecture integrating linguistic cues with vector storage to improve long-term dialogue performance.

Findings

Up to 18% improvement in factual recall and discourse coherence.

Enhanced robustness and interpretability demonstrated through ablation and human evaluations.

Effective integration of dependency parsing, discourse relations, and coreference resolution.

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.