Research paper introducing dropout as a regularization technique for neural networks.
Paper: Dropout - A Simple Way to Prevent Neural Networks from Overfitting.pdf
Related
- Deep Residual Learning for Image Recognition — both are foundational deep-learning generalization/training techniques; note ResNet itself does not use dropout (it relies on batch normalization instead).
- Layer Normalization — general “regularizing/normalizing deep network training” connection; the two techniques (random unit dropping vs. activation normalization) are not directly dependent on each other and were published two years apart (2014 vs. 2016).