TY  - RPRT
TI  - PRIME: Training Free Proactive Reasoning via Iterative Memory Evolution for User-Centric Agent
AU  - Prince Zizhuang Wang
AU  - Shuli Jiang
PY  - 2026
UR  - https://arxiv.org/abs/2604.07645
ID  - 2604.07645
ER  -