Limitation on the computational resources is an issue faced in machine learning. In response, random neural networks and non-linear projections can be used, e.g. to train single-layer neural networks very fast (often called extreme learning machines in recent literature) and yet to obtain satisfactory prediction performances. I analysed their behaviour when the number of neurons becomes very large. In particular, I shown that such models can be formulated in terms of a kernel for which an analytical form exists.
- Frénay, B., Verleysen M. Parameter-insensitive kernel in extreme learning for non-linear support vector regression. Neurocomputing, 74(16), 2526–2531, 2011.
- Frénay, B., Verleysen M. Using SVMs with randomised feature spaces: an extreme learning approach. In Proc. ESANN 2010, 315–320.