Zeyuan publishes on ICASSP

Our paper entitled On Optimization of Poles for Adaptive Fourier Decomposition-inspired Neural Layers is accpeted by ICASSP 2026, one of the top conferences in acoustics, speech, and signal processing. This work presents a novel approach to optimize pole locations and network hyperparameters for adaptive Fourier decomposition-inspired neural operators to improve its accuracy in solving partial differential equations.