Natural Language Generation for a Smart Biology Textbook

In this demo paper we describe the natural language generation component of an electronic textbook application, called Inquire1. Inquire interacts with a knowledge base which encodes information from a biology textbook. The application includes a question-understanding module which allows students to ask questions about the contents of the book, and a questionanswering module which retrieves the corresponding answer from the knowledge base. The task of the natural language generation module is to present specific parts of the answer in English. Our current generation pipeline handles inputs that describe the biological functions of entities, the steps of biological processes, and the spatial relations between parts of entities. Our ultimate goal is to generate paragraphlength texts from arbitrary paths in the knowledge base. We describe here the natural language generation pipeline and demonstrate the inputs and generated texts. In the demo presentation we will show the textbook application and the knowledge base authoring environment, and provide an opportunity to interact with the system.