Tag
This paper introduces Wave Function Backpropagation (WFB), a wave-parameterized learning formulation that explicitly models temporal intervals to improve neural network performance in trajectory prediction, achieving up to 20.4% reduction in average displacement error.