kdehumor at semeval-2020 task 7: a neural network model for detecting funniness in dataset humicroedit

TL;DR

This paper presents a neural network approach using BiLSTMs and pre-trained sentence embeddings to detect humor in news headlines, specifically for the SemEval-2020 Task 7 dataset.

Contribution

The paper introduces a deep neural network model combining BiLSTMs and sentence embeddings for humor detection in news headlines.

Findings

The model effectively captures humor cues in headlines.

Pre-trained sentence embeddings improve detection accuracy.

Component analysis highlights the importance of each architecture part.

Abstract

This paper describes our contribution to SemEval-2020 Task 7: Assessing Humor in Edited News Headlines. Here we present a method based on a deep neural network. In recent years, quite some attention has been devoted to humor production and perception. Our team KdeHumor employs recurrent neural network models including Bi-Directional LSTMs (BiLSTMs). Moreover, we utilize the state-of-the-art pre-trained sentence embedding techniques. We analyze the performance of our method and demonstrate the contribution of each component of our architecture.