Ensemble-based decoders can effectively capture the predictive uncertainty in deep learning-based video compression models, leading to improved rate-distortion performance.
