Personality-dependent content selection in natural language generation systems

Ricelli M S Ramos,Ivandré Paraboni,Danielle S Monteiro

doi:10.1186/s13173-020-00096-1

Ricelli M S Ramos, Ivandré Paraboni + Show 1 more

Open Access

https://doi.org/10.1186/s13173-020-00096-1

Copy DOI

Abstract

This paper focuses on the computer side of human-computer interaction through natural language, which is the domain of natural language generation (NLG) studies. From a given (usually non-linguistic) input, NLG systems will in principle generate the same fixed text as an output and in order to attain more natural or human-like interaction will often resort to a wide range of strategies for stylistic variation. Among these, the use of computational models of human personality has emerged as a popular alternative in the field and will be the focus of the present work as well. More specifically, the present study describes two machine learning experiments to establish possible relations between personality and content selection (as opposed to the more well-documented relation between personality and surface realisation), and it is, to the best of our knowledge, the first of its kind to address this issue at both macro and micro planning levels, which may arguably pave the way for the future development of more robust personality-dependent systems of this kind.

Highlights

Natural language generation (NLG) systems produce text from non-linguistic input and are central to the development of realistic, psychologically plausible humancomputer communication that does not resort to pre-defined or ‘canned’ text
Our work will disregard the issue of which scene objects are to be selected and will focus instead on the choices of facts about these objects the present experiment should be more appropriately described as a first step towards more comprehensive document planning studies rather than a fully functional content selection (CS) module
Both models were built using linear support vector machine (SVM) with optimal parameter values obtained by performing grid search on the training dataset

Summary

Introduction

Natural language generation (NLG) systems produce text from (usually) non-linguistic input and are central to the development of realistic, psychologically plausible humancomputer communication that does not resort to pre-defined or ‘canned’ text. One possible way of adding human variation to the output descriptions generated by a REG algorithm without resorting to a large among of linguistic examples as training data is by assuming that personality may play a role in the content selection of referring expressions as well.

Objectives

Results

Conclusion