Natural Language Generation (NLG)
Date:
The goal of this seminar was to explore the classical three stage architecture along with it’s several tasks in detail. A data driven news generation system for automated journalism was also explored.
From a practical perspective the most common architecture in the present day applied NLG systems is the a modularized pipeline based architecture. It is a classical NLG architecture with three stages that define the complete pipeline. Each stage can have one or more tasks. There is a general consensus in the natural language community that there are six basic tasks that needs to be carried out for transforming input data into output text in some natural language. A stage wise separation of these tasks for the classical three stage architecture can be seen in the following figure.
