Faria, P. (2020). Compreendendo a modelagem computacional de aquisição da linguagem. Veredas – Revista de Estudos Linguísticos, v.24, n.1.

Abstract

It is common to hear language acquisition researchers complain that computational models are too unfamiliar and relatively inaccessible. However, a fruitful exchange of ideas and mutual perceptions on models is important for the advance of acquisition theories. Taking as an example problem the distributional learning of syntactic categories, we present an example of a process which starts with a description of the acquisition problem, moves to its “translation” into a modeling problem, and ends with the model itself. In this route, we become aware of how modeling imposes that assumptions, simplifications, and arbitrary decisions, are made explicit. Understanding the process which leads us to computational models will enable us to, as a community, better evaluate them and, more than just pointing out limitations, also point out directions for further development, or for fully exploring their potential, through the formulation of novel experiments to answer novel questions. (paper in Portuguese)

Keywords

Language Acquisition, Computational Modeling, Distributional Learning, Methodology.