Transformer based ensemble for emotion detection

TL;DR

This paper presents an ensemble of ELECTRA and BERT models for emotion detection in essays, achieving an F1 score of 62.76%, contributing to the WASSA 2022 shared task.

Contribution

It introduces a novel ensemble approach combining ELECTRA and BERT for emotion classification in text, with publicly available code and results.

Findings

Achieved an F1 score of 62.76% on the WASSA 2022 task.

Demonstrated the effectiveness of ensemble models for emotion detection.

Provided open access to code and experimental setup.

Abstract

Detecting emotions in languages is important to accomplish a complete interaction between humans and machines. This paper describes our contribution to the WASSA 2022 shared task which handles this crucial task of emotion detection. We have to identify the following emotions: sadness, surprise, neutral, anger, fear, disgust, joy based on a given essay text. We are using an ensemble of ELECTRA and BERT models to tackle this problem achieving an F1 score of $62.76\%$. Our codebase (https://bit.ly/WASSA_shared_task) and our WandB project (https://wandb.ai/acl_wassa_pictxmanipal/acl_wassa) is publicly available.