The phrase "a set of kernels" is correct and usable in written English.
It can be used when referring to a collection or group of kernels, often in contexts related to computing, machine learning, or cooking.
Example: "In machine learning, a set of kernels can be used to improve the performance of support vector machines."
Alternatives: "a collection of kernels" or "a group of kernels".
A set of kernels is called characteristic kernels, introduced in [4, 5] gives an RKHS for which probabilities have unique images.
By using all available data encoded into a set of kernels, MKL classifiers most frequently outperform a single kernel classifier constructed for one type of data.
The analysis of kernels can also help construct new interventions by putting together a set of kernels that all appear relevant and useful for new problems.
In the following, we define a set of kernels for fragmentation trees that will allow us to transfer the power of the fragmentation tree approach to the kernel-based learning algorithms for molecular fingerprint prediction and metabolite identification.
For all the three algorithms, the input will be a set of kernels K = { K k | K k ∈ ℝ n × n, k = 1, …, q } computed from n data points.
The function K denotes any of a set of kernel functions.
This design consists of a set of kernel functions and related parameters describing the message.
For example, in multiple kernel learning, a kernel is constructed from each view and a set of kernel coefficients are learned to obtain an optimal combined kernel matrix.
The classifier effectively learns a folding characteristic for each bit, which is realized as the weighted sum of a set of kernel functions.
For a scale-space representation based on a multi-dimensional scale parameter, one may also consider a weaker requirement of rotational invariance at the level of a family of kernels, for example regarding a set of elongated kernels with different orientations in image space.
However, the best possible approximation of the local kernels as linear interpolations of a set of reference kernels (for a certain error measurement) is obtained when no spatial constraints are imposed to the linear coefficients.
