Modeling Language Model Structure

af Mary R. Buchanan

Neural language models are probabilistic models of text parameterized by neural networks. They are widely applicable to applications with outputs consisting of discrete sequences, such as document summarization, question answering, and image captioning. The minimal assumptions about data enable advancements in language modeling to drive improvements across a diverse array of applications. In natural language, structures are both pervasive and essential. For example, a book is organized into chap... (Læs mere)