Modeling morphology with Linear Discriminative Learning: considerations and design choices
It addresses incremental learning and representation choices for linguists and computational modelers, but is incremental in refining an existing approach.
This study tackled methodological questions in modeling German noun inflection using Linear Discriminative Learning, showing that the model provides excellent memory for known words but limited performance for unseen data, aligning with semi-productivity and native speaker generalization.
This study addresses a series of methodological questions that arise when modeling inflectional morphology with Linear Discriminative Learning. Taking the semi-productive German noun system as example, we illustrate how decisions made about the representation of form and meaning influence model performance. We clarify that for modeling frequency effects in learning, it is essential to make use of incremental learning rather than the endstate of learning. We also discuss how the model can be set up to approximate the learning of inflected words in context. In addition, we illustrate how in this approach the wug task can be modeled in considerable detail. In general, the model provides an excellent memory for known words, but appropriately shows more limited performance for unseen data, in line with the semi-productivity of German noun inflection and generalization performance of native German speakers.