HypoGen: Hyperbole Generation with Commonsense and Counterfactual Knowledge

TL;DR

This paper introduces HypoGen, a novel method for sentence-level hyperbole generation that leverages commonsense and counterfactual knowledge, achieving creative and high-quality exaggerations through a combination of pattern-based generation and neural ranking.

Contribution

The paper presents a new approach to hyperbole generation using semantic inference models and a ranking system, addressing the scarcity of computational hyperbole research.

Findings

Generated hyperboles have high success and intensity scores.

The method outperforms baselines in automatic and human evaluations.

Leveraging commonsense and counterfactual inference improves hyperbole quality.

Abstract

A hyperbole is an intentional and creative exaggeration not to be taken literally. Despite its ubiquity in daily life, the computational explorations of hyperboles are scarce. In this paper, we tackle the under-explored and challenging task: sentence-level hyperbole generation. We start with a representative syntactic pattern for intensification and systematically study the semantic (commonsense and counterfactual) relationships between each component in such hyperboles. Next, we leverage the COMeT and reverse COMeT models to do commonsense and counterfactual inference. We then generate multiple hyperbole candidates based on our findings from the pattern, and train neural classifiers to rank and select high-quality hyperboles. Automatic and human evaluations show that our generation method is able to generate hyperboles creatively with high success rate and intensity scores.