The most convincing models of human grammar acquisition to date are supervised, in the sense that they learn from pairs of strings and meaning representations (Siskind, 1996; Villavicencio, 2002; Villavicencio, 2011; Buttery, 2004; Buttery, 2006; Kwiatkowski et al., 2012). Although the principles by which such models learn are quite general, the datasets they have been applied to have unavoidably been somewhat target-language-specific, and are also limited to discourse-external world-state-related content, contrary to the observations of (Tomasello, 2001) concerning the central role of common ground and grounding in interpersonal interaction.
|Title of host publication||Proceedings of the Workshop on Computational Models of Language Acquisition and Loss|
|Place of Publication||Stroudsburg, PA, USA|
|Publisher||Association for Computational Linguistics|
|Number of pages||1|
|Publication status||Published - 2012|