Adaptive scheduling for adaptive sampling in POS taggers construction

We introduce an adaptive scheduling for adaptive sampling as a novel way of machine learning in the construction of part-of-speech taggers. The goal is to speed up the training on large data sets, without significant loss of performance with regard to an optimal configuration. In contrast to previous methods using a random, fixed or regularly rising spacing between the instances, ours analyzes the shape of the learning curve geometrically in conjunction with a functional model to increase or decrease it …

construction contrast cs.ai cs.cl cs.lg data data sets loss machine machine learning novel part part-of-speech performance regard sampling scheduling speech speed training