Tag
This paper proposes MVG-KAN, a multi-view model integrating periodic-residual decomposition, a Geo-Wind Graph for wind-aware spatial dependencies, and a temporal KAN head for PM2.5 forecasting, achieving MAE 14.09 on Beijing data.
This paper presents an alternative architecture for LLMs using Radial Basis Function (RBF) networks that eliminates deep neural networks and finds the global optimum in closed form, requiring no iterative training. It also reviews other non-DNN methods like KANs and k-NN retrieval, with a case study demonstrating increased explainability and faster training.