An Overview of Natural Language Generation Systems Evaluation

Abstract

The evaluation of natural language generation systems has been an eyes-catching topic of research that is continuously challenging researchers in the effort towards achieving the performance and quality of computational linguistic systems. Although this research topic keeps drawing researchers attention, there is still no agreed benchmark on what should be evaluated and how to evaluate them. This paper analyzed and compiled the evaluations of some well-known natural language generation systems and categorized their evaluation methods according their relevancies and similarities.

Topics

1 Figures and Tables

Download Full PDF Version (Non-Commercial Use)