Our paper “FLAME: Flow Enhanced Legendre Memory Models for General Time Series Forecasting” was accepted as a poster at NeurIPS 2026. Congratulations to Xingjian Wu.