Duluth at SemEval-2020 Task 7: Using Surprise as a Key to Unlock Humorous Headlines

TL;DR

This paper explores using pretrained transformer models and a contrastive approach based on surprise to assess humor in news headlines, achieving competitive results in SemEval-2020.

Contribution

It introduces a novel contrastive method leveraging surprise detection with transformer models for humor assessment in headlines.

Findings

Achieved 0.531 RMSE in Subtask 1, ranking 11th

Achieved 0.632 accuracy in Subtask 2, ranking 9th

Demonstrated effectiveness of surprise-based contrastive approach

Abstract

We use pretrained transformer-based language models in SemEval-2020 Task 7: Assessing the Funniness of Edited News Headlines. Inspired by the incongruity theory of humor, we use a contrastive approach to capture the surprise in the edited headlines. In the official evaluation, our system gets 0.531 RMSE in Subtask 1, 11th among 49 submissions. In Subtask 2, our system gets 0.632 accuracy, 9th among 32 submissions.