Wearable accelerometers are widely used as an ecologically valid and scalable solution for long-term at-home sleep monitoring in both clinical research and care. The levels of accuracy with accelerometry-based sleep outcomes, however, have not improved much for the last 40 years. In this study, we applied deep-learning convolutional neural network (CNN) models to this task and demonstrated that this new model outperformed existing sleep algorithms in classifying sleep-wake and estimating sleep outcomes based on wrist-worn accelerometry. We further showed that, using a domain adversarial technique, this model generalized well to another dataset based on different wearable devices and activity counts, achieving an accuracy of 80.1 (sensitivity 84% and specificity 58%). Compared to the commonly used sleep algorithms, this model, termed DAsleepCNN, resulted in the smallest error in wake after sleep onset (WASO) and sleep efficiency (SE) outcomes. As wearable-based digital health technologies are increasingly used in clinical trials and care, improving the accuracy and generalizability of sleep algorithms for wrist-worn wearable data are of utmost importance. We here demonstrated that domain adversarial convolutional neural networks can improve the overall accuracy, especially the specificity, of sleep-wake classification using wrist-worn accelerometer data, substantiating its use as a scalable and valid approach for sleep outcome assessment in real-life.