Study optimizes deep learning models for sleep stage classification.
problem Time-consuming and inconsistent manual sleep stage scoring.
method Investigated architectural choices in encoder-predictor architectures for polysomnography recordings.
result Robust architectures improve sleep stage classification performance.