Genre Editor
Designing a generative word processor
This work was done as a course project for Stony Brook University’s CSE 518 Foundations of Human Computer Interaction, Fall 2020. The aim of this project was to design and evaluate a word processor with generative capabilities in order to aid the writing of short fiction.
Abstract:
Recent advances in natural language processing have been adopted widely in various fields and products. Natural language generation in particular has been used to display real-time predictive text while typing on QWERTY keypads, search engines, and even email. In this project, we utilize recent progress on topic specific language generation to develop a proof of concept implementation of a word processor with predictive capabilities. We also present a heuristic evaluation of our system, and propose modifications based on user feedback we gathered.