This paper introduces a novel approach called "Auto-DSP" that aims to automatically learn optimal update rules for adaptive filtering algorithms, particularly in the context of acoustic echo cancellation.