Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note2 min read

Hierarchical Graph Memory for LLM Agents with Path-level Localization and Rewrite

Source

arXiv

arxiv.org/abs/2608.05095v1

Read the source

Source

Plain-Language Brief

The authors propose HiGram, a hierarchical graph memory framework for long-term reasoning agents. This framework aims to efficiently update memories over time by organizing them into a coarse-to-fine architecture, reducing irrelevant information during retrieval and improving answer quality and token efficiency in conversational question answering.

What the Document Says

  • The authors propose HiGram, an evolving hierarchical graph memory framework.
  • It organizes the memory into a coarse-to-fine architecture.
  • Reduces irrelevant information during retrieval.
  • Improves answer quality and token efficiency in conversational question answering.
  • Demonstrates substantial improvements over baselines.

Key Concepts Mentioned

  • graph memory
  • hierarchical architecture
  • long-term reasoning
  • retrieval
  • answer quality
  • token efficiency

Claims Made by the Document

  • Existing methods store all memories in a flat graph.
  • Accumulated historical memories can introduce irrelevant contexts and increase the cost of evidence selection during retrieval.
  • They typically update memory units independently, requiring repeated unit-wise rewrite to cover related changes.
  • Our method demonstrates substantial improvements over baselines in answer quality and token efficiency.

Evidence Used

  • The authors state that agents for long term reasoning require a memory that can be efficiently and effectively updated over time.
  • They mention that graph memory has been adopted to offer structural organization for multi-hop retrieval and reasoning.
  • The authors explain that existing methods store all memories in a flat graph and accumulated historical memories can introduce irrelevant contexts.

Methods / Approach

  • Hierarchical graph memory
  • MicroGraph-based path-level localization
  • Coordinated rewriting method

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Agents for long term reasoning require a memory that can be efficiently and effectively updated over time, as new facts and external feedback continue to arrive. Recently, graph memory has been adopted to offer structural organization for multi-hop retrieval and reasoning. However, existing methods store all memories in a flat graph, and accumulated historical memories can introduce irrelevant contexts and increase the cost of evidence selection during retrieval. Moreover, they typically update memory units independently, requiring repeated unit-wise rewrite to cover related changes. To address these issues, we propose HiGram, an evolving hierarchical graph memory framework with path-level localization and rewriting. Specifically, we first propose a hierarchical graph memory, which organizes the memory into coarse-to-fine architecture composed of upper-level nodes and MemoryUnits, thereby reducing the amount of irrelevant information during retrieval. We further propose MicroGraph-based path-level localization, which leverages query and update conditioned MicroGraphs to identify support subgraph and evidence path before rewrite. Finally, we propose a coordinated rewriting method that jointly revises intra-unit memory and inter-unit dependencies, enable valid dependency structures updating in the localized evidence path. Experiments on benchmarks for long-term conversational question answering and conflict-aware memory evaluation demonstrate that our method demonstrate substantial improvements over baselines in answer quality and token efficiency. Besides, our method improves answer accuracy and query-valid evidence selection under dynamic, static, and conditional conflicts.