Hippocampus: An Efficient and Scalable Memory Module for Agentic AI

2026-02-14

AI Summary

This work addresses critical bottlenecks in existing AI agent memory systems—namely, retrieval latency and storage scalability. We propose Hippocampus, a novel memory management system that introduces dynamic wavelet matrices (DWM) into agent memory for the first time. By integrating compact binary semantic signatures with lossless token-ID streams, Hippocampus enables efficient joint indexing and content reconstruction directly in the compressed domain. The approach achieves linear scalability, low-latency retrieval, and high precision, maintaining competitive accuracy on the LoCoMo and LongMemEval benchmarks while reducing end-to-end retrieval latency by up to 31× and decreasing per-query token footprint by as much as 14×.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Multiagent Systems: Agent Communication

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

Agentic AI require persistent memory to store user-specific histories beyond the limited context window of LLMs. Existing memory systems use dense vector databases or knowledge-graph traversal (or hybrid), incurring high retrieval latency and poor storage scalability. We introduce Hippocampus, an agentic memory management system that uses compact binary signatures for semantic search and lossless token-ID streams for exact content reconstruction. Its core is a Dynamic Wavelet Matrix (DWM) that compresses and co-indexes both streams to support ultra-fast search in the compressed domain, thus avoiding costly dense-vector or graph computations. This design scales linearly with memory size, making it suitable for long-horizon agentic deployments. Empirically, our evaluation shows that Hippocampus reduces end-to-end retrieval latency by up to 31$\times$ and cuts per-query token footprint by up to 14$\times$, while maintaining accuracy on both LoCoMo and LongMemEval benchmarks.

Problem

• Agentic AI
• memory systems
• retrieval latency
• storage scalability
• persistent memory

Innovation

• Dynamic Wavelet Matrix
• binary semantic signatures
• lossless token-ID streams
• compressed-domain search
• agentic memory