Explainable Volterra Networks Bring Transparent Nonlinear Dynamics to Time Series Forecasting
A new neural architecture converts the classical Volterra series into learnable coefficient tensors, recovering known chaotic dynamics, matching transformer physics, ...
