TY - RPRT TI - A Multi-Agent Approach to Validate and Refine LLM-Generated Personalized Math Problems AU - Fareya Ikram AU - Nischal Ashok Kumar AU - Junyang Lu AU - Hunter McNichols AU - Candace Walkington AU - Neil Heffernan AU - Andrew S. Lan PY - 2026 UR - https://arxiv.org/abs/2604.05160 ID - 2604.05160 ER -