MAPLE: A Sub-Agent Architecture for Memory, Learning, and Personalization in Agentic AI Systems

TL;DR

This paper introduces MAPLE, a modular architecture for agentic AI systems that distinctly separates memory, learning, and personalization to improve adaptability and personalization effectiveness.

Contribution

It proposes a novel decomposition of agent architecture into three specialized sub-agents, each optimized independently, enhancing personalization and learning capabilities.

Findings

Achieved 14.6% improvement in personalization score

Increased trait incorporation rate from 45% to 75%

Demonstrated effective independent optimization of components

Abstract

Large language model (LLM) agents have emerged as powerful tools for complex tasks, yet their ability to adapt to individual users remains fundamentally limited. We argue this limitation stems from a critical architectural conflation: current systems treat memory, learning, and personalization as a unified capability rather than three distinct mechanisms requiring different infrastructure, operating on different timescales, and benefiting from independent optimization. We propose MAPLE (Memory-Adaptive Personalized LEarning), a principled decomposition where Memory handles storage and retrieval infrastructure; Learning extracts intelligence from accumulated interactions asynchronously; and Personalization applies learned knowledge in real-time within finite context budgets. Each component operates as a dedicated sub-agent with specialized tooling and well-defined interfaces. Experimental evaluation on the MAPLE-Personas benchmark demonstrates that our decomposition achieves a 14.6% improvement in personalization score compared to a stateless baseline (p < 0.01, Cohen's d = 0.95) and increases trait incorporation rate from 45% to 75% -- enabling agents that genuinely learn and adapt.