RhyMix: A Lightweight Adaptive Multi-Rhythm Network for Long-Term Time Series Forecasting
RhyMix is a lightweight hybrid neural architecture designed to forecast complex time series by combining two paths: one capturing explicit seasonal cycles and another using multi-scale convolutions for local patterns. It uses adaptive gating mechanisms to dynamically fuse these paths, achieving state-of-the-art performance with linear complexity and low latency suitable for edge devices.