EUREKA: EUphemism Recognition Enhanced through Knn-based methods and Augmentation

TL;DR

EUREKA is an ensemble approach that improves euphemism detection by correcting dataset labels, expanding data with EuphAug, and utilizing kNN-based representations, achieving state-of-the-art results.

Contribution

It introduces a novel ensemble method with data augmentation and representation techniques for enhanced euphemism detection.

Findings

Achieved a macro F1 score of 0.881 on the Euphemism Detection Shared Task.

Successfully curated an expanded corpus called EuphAug.

Outperformed previous methods on the public leaderboard.

Abstract

We introduce EUREKA, an ensemble-based approach for performing automatic euphemism detection. We (1) identify and correct potentially mislabelled rows in the dataset, (2) curate an expanded corpus called EuphAug, (3) leverage model representations of Potentially Euphemistic Terms (PETs), and (4) explore using representations of semantically close sentences to aid in classification. Using our augmented dataset and kNN-based methods, EUREKA was able to achieve state-of-the-art results on the public leaderboard of the Euphemism Detection Shared Task, ranking first with a macro F1 score of 0.881. Our code is available at https://github.com/sedrickkeh/EUREKA.