Languages as hyperplanes: grammatical inference with string kernels

TY - JOUR

T1 - Languages as hyperplanes: grammatical inference with string kernels

AU - Clark, Alexander

AU - Costa Florencio, Christophe

AU - Watkins, Chris

N1 - 10.1007/s10994-010-5218-3

PY - 2011/3

Y1 - 2011/3

N2 - Using string kernels, languages can be represented as hyperplanes in a high dimensional feature space. We discuss the language-theoretic properties of this formalism with particular reference to the implicit feature maps defined by string kernels, considering the expressive power of the formalism, its closure properties and its relationship to other formalisms. We present a new family of grammatical inference algorithms based on this idea. We demonstrate that some mildly context-sensitive languages can be represented in this way and that it is possible to efficiently learn these using kernel PCA. We experimentally demonstrate the effectiveness of this approach on some standard examples of context-sensitive languages using small synthetic data sets.

AB - Using string kernels, languages can be represented as hyperplanes in a high dimensional feature space. We discuss the language-theoretic properties of this formalism with particular reference to the implicit feature maps defined by string kernels, considering the expressive power of the formalism, its closure properties and its relationship to other formalisms. We present a new family of grammatical inference algorithms based on this idea. We demonstrate that some mildly context-sensitive languages can be represented in this way and that it is possible to efficiently learn these using kernel PCA. We experimentally demonstrate the effectiveness of this approach on some standard examples of context-sensitive languages using small synthetic data sets.

U2 - 10.1007/s10994-010-5218-3

DO - 10.1007/s10994-010-5218-3

M3 - Article

SN - 0885-6125

VL - 82

SP - 351

EP - 373

JO - MACHINE LEARNING

JF - MACHINE LEARNING

IS - 3

ER -