PersonaTree: Structured Lifecycle Memory for Person Understanding in LLM Agents
2026-06-03 • Computation and Language
Computation and Language
AI summaryⓘ
The authors focus on how AI agents remember and understand people over long interactions by organizing memories into a structured 'persona tree.' This tree groups evidence from interactions into higher-level patterns and stable claims about a person, making the AI's understanding clearer and more precise. They created a system called PersonaTree that carefully updates this tree and fetches only the relevant memory when needed. Tests show PersonaTree performs very well compared to other methods in tasks involving understanding people and remembering information over time.
LLM agentsmemory representationspersona understandingschema formationPersonaTreestructured memoryevidence abstractionperson claimssupport path retrievalpersistent memory benchmarks
Authors
Yubo Hou, Jingwei Song, Hongbo Zhang, Zhisheng Chen, Bang Xiao, Tao Wan, Zengchang Qin
Abstract
Persistent LLM agents require memory representations that make the formation of person understanding explicit across long term interaction. Existing agent memory methods emphasize information retention and retrieval, yet give limited account of how accumulated interaction evidence is abstracted into person understanding. We view this process as schema formation, where situated evidence is abstracted into reusable patterns and stable person level claims. We introduce PersonaTree, a structured lifecycle memory framework that realizes this view as a three level persona tree with explicit support paths from evidence to claims. PersonaTree maintains the tree through conservative writing, confidence guided consolidation, and query conditioned path retrieval, returning only the evidence depth required by each query. Across six person understanding and persistent memory benchmarks with three answer backbones, PersonaTree ranks first in 12 of 18 compact scores and reaches the top two in 16 settings. Ablations show that hierarchy improves abstract person understanding on KnowMe, while support path retrieval improves RealPref alignment under a comparable context budget.