1. Adaptive dropout for training deep neural networks;Ba,2013
2. Training with noise is equivalent to Tikhonov regularization;Bishop;Neural Computation,1995
3. Bouthillier, X., Konda, K., Vincent, P., & Memisevic, R. (2015). Dropout as data augmentation. arXiv preprint arXiv:1506.08700.
4. A review on neural networks with random weights;Cao;Neurocomputing,2018
5. Natural language processing (almost) from scratch;Collobert;Journal of Machine Learning Research (JMLR),2011