VikingMem: A Memory Base Management System for Stateful LLM-based Applications

2026-05-28

AI Summary

Large language models are constrained by limited context windows, hindering their ability to support long-term, stateful interactive applications; existing memory mechanisms suffer from poor generalization and incomplete retrieval. This work proposes the Memory Base paradigm, introducing a novel memory architecture grounded in three core principles: selective extraction, endogenous state evolution, and universal abstraction. By leveraging event-entity abstraction, the framework enables end-to-end memory management. Integrated with the VikingDB vector engine, the system incorporates event-centric extraction, thematic timeline compression, temporally weighted recall, and dynamic entity updating. Evaluated on long-term memory benchmarks, it achieves up to a 30% improvement in retrieval performance over baseline methods while maintaining low latency, demonstrating broad applicability across domains such as education, recommendation systems, and intelligent agents.

Abstract

Large Language Models have revolutionized interactive applications; however, their finite context windows pose a critical data management challenge for maintaining stateful, long-term interactions. Existing memory approaches often rely on simplistic extraction methods that lead to incomplete memories or use rigid, single-purpose memory extraction prompts tailored to a single use case, such as chatbots. Consequently, they lack generalizability and perform poorly across diverse downstream tasks. To bridge this gap, we introduce the Memory Base, a novel data management paradigm for managing the persistent state of long-term interactions. It is characterized by three core principles: selective extraction of high-value memories from raw information streams; inherent statefulness and evolution, where memory content is progressively summarized, corrected, and temporally weighted to prioritize recent interactions; and a generalizable abstraction paradigm designed for robust transferability across diverse applications, including education, recommendation, and agent memory. Building on this foundation, we present VikingMem, an end-to-end Memory Base Management System implemented on the VikingDB vector engine. VikingMem materializes this paradigm through interconnected event and entity abstractions. It features event-centric memory extraction to selectively handle complex information streams, while entities are dynamically updated by events to achieve stateful evolution. Using temporal compression via a topic-wise timeline and time-weighted recall, the system progressively produces high-level summary memories, prioritizes recent items, and compresses and fades older ones. Extensive evaluations on long-term memory benchmarks demonstrate that VikingMem outperformes baselines by up to 30% in memory retrieval effectiveness while maintaining the low latency essential for interactive applications.

Problem

• stateful LLM
• memory management
• long-term interaction
• context window limitation
• generalizable memory

Innovation

• Memory Base
• stateful LLM
• event-entity abstraction
• temporal compression
• generalizable memory management