Any linear model can be turned into a non-linear model by applying the kernel trick to the model—replacing its features (predictors) by a kernel function. In other words, the kernel implicitly transforms our dataset into higher dimensions. The kernel trick leverages the fact that it is often easier to separate the instances in more dimensions. Algorithms capable of operating with kernels include the kernel perceptron, SVMs, Gaussian processes, PCA, canonical correlation analysis, ridge regression, spectral clustering, linear adaptive filters, and many others.