ABSTRACT
In particular, one should look to machine learning, the subfield of artificial intelligence (AI) concerned with computational approaches to learning - i.e., with processes that lead to improved performance over time. This chapter begins by reviewing four basic learning tasks that have been the focus of the machine learning work, describing each in terms of a common framework. It also considers the tasks in roughly historical order, based on the periods at which they first drew the major attention of machine learning researchers. Then, describes grammar learning, comparing and contrasting it with other learning tasks. Next reviews some earlier computational models of grammar learning and considers some drawbacks of these models. Finally, it outlines a new approach to the modeling of language acquisition that will overcome these limitations.
