AriadneMem: Threading the Maze of Lifelong Memory for LLM Agents

2025-11-16 · Source: cs.CL updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Robotics & Autonomous Systems · Depth: Expert, extended

Summary

AriadneMem is a novel structured memory system designed for long-horizon LLM agents, specifically addressing challenges of "disconnected evidence" and "state updates" within fixed context budgets. It employs a decoupled two-phase pipeline: an offline construction phase uses "entropy-aware gating" to filter noise and "conflict-aware coarsening" to build an evolutionary graph that explicitly preserves state transitions as temporal edges. The online reasoning phase then executes "algorithmic bridge discovery" to reconstruct logical paths between retrieved facts, followed by "single-call topology-aware synthesis," eliminating expensive iterative planning. This approach significantly improves performance, achieving a 15.2% increase in Multi-Hop F1 and a 9.0% gain in Average F1 on LoCoMo experiments with GPT-4o, while crucially reducing total runtime by 77.8% using only 497 context tokens.

Key takeaway

AriadneMem is a novel structured memory system for LLM agents that resolves disconnected evidence and state update challenges in long-term dialogues. It employs offline entropy-aware gating and conflict-aware coarsening, alongside online algorithmic bridge discovery and single-call topology-aware synthesis. This approach boosts Multi-Hop F1 by 15.2% and Average F1 by 9.0% on LoCoMo (GPT-4o), while cutting total runtime by 77.8% using only 497 context tokens, making complex reasoning efficient for lifelong agents.

Topics

• LLM Agents
• Structured Memory Systems
• Graph-based Memory
• Multi-Hop Reasoning
• Conflict-Aware Coarsening

Code references

• LLM-VLM-GSL/AriadneMem
• mem0ai/mem0
• snap-research/locomo

Best for: AI Scientist, Research Scientist, NLP Engineer, AI Researcher, AI Engineer, Machine Learning Engineer

Related on AIssential

• StructMem: Structured Memory for Long-Horizon Behavior in LLMs — StructMem offers a hierarchical memory for LLMs, balancing efficiency and relational structure for long-term conversations.
• TOKI: A Bitemporal Operator Algebra for Contradiction Resolution in LLM-Agent Persistent Memory — Contradiction resolution in LLM agent memory is write-time concurrency control, requiring explicit isolation and provenance.
• LongMINT: Evaluating Memory under Multi-Target Interference in Long-Horizon Agent Systems — Current memory-augmented agents struggle with interference and multi-target reasoning in long, dynamic contexts.
• Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents — LLM agent memory should be actively reconstructed and dynamically adapted to reasoning context, not statically retrieved.
• The cognitive companion: a lightweight parallel monitoring architecture for detecting and recovering from reasoning degradation in LLM agents — A parallel monitoring architecture can detect and mitigate LLM agent reasoning degradation with minimal or zero overhead.
• A Practical Guide to Memory for Autonomous LLM Agents — Memory architecture, particularly its management, is more critical to LLM agent performance than model choice.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.CL updates on arXiv.org.