It is reported as one of the techniques but I am not able to get a reference.
"Jittering" refers to the idea of adding small amounts of noise to your data to generate new, artificially corrupted examples. Theoretically, it is known to have a regularizing effect on the solution. Some of the most important references in this respect are:
[1] Koistinen, Petri, and Lasse Holmstrom. "Kernel regression and backpropagation training with noise." Neural Networks, 1991. 1991 IEEE International Joint Conference on. IEEE, 1991.
[2] Holmstrom, Lasse, and Petri Koistinen. "Using additive noise in back-propagation training." Neural Networks, IEEE Transactions on 3.1 (1992): 24-38.
[3] Bishop, Chris M. "Training with noise is equivalent to Tikhonov regularization." Neural computation 7.1 (1995): 108-116.
[4] An, Guozhong. "The effects of adding noise during backpropagation training on a generalization performance." Neural Computation 8.3 (1996): 643-674.
Recently, the idea of adding artificial noise to the data has been used extensively in the field of deep learning, especially for the "Stacked Denoising Autoencoder":
[5] Vincent, Pascal, et al. "Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion." The Journal of Machine Learning Research 9999 (2010): 3371-3408.
There is also a nice, less formal explanation of jittering here:
ftp://ftp.sas.com/pub/neural/FAQ3.html#A_jitter
I hope this helps.
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining