VISU at WASSA 2023 Shared Task: Detecting Emotions in Reaction to News Stories Leveraging BERT and Stacked Embeddings

TL;DR

This paper presents the VISU system that employs deep learning models with stacked embeddings and tailored preprocessing to detect emotions in reactions to news stories, addressing challenges of complex dialogue understanding.

Contribution

The study introduces a novel combination of static and contextual embeddings with deep learning models for emotion detection in news reaction essays.

Findings

Achieved a Macro F1-Score of 0.2717 in the WASSA 2023 shared task.

Demonstrated effectiveness of stacked embeddings and tailored preprocessing for small, imbalanced datasets.

Ranked tenth among participating systems.

Abstract

Our system, VISU, participated in the WASSA 2023 Shared Task (3) of Emotion Classification from essays written in reaction to news articles. Emotion detection from complex dialogues is challenging and often requires context/domain understanding. Therefore in this research, we have focused on developing deep learning (DL) models using the combination of word embedding representations with tailored prepossessing strategies to capture the nuances of emotions expressed. Our experiments used static and contextual embeddings (individual and stacked) with Bidirectional Long short-term memory (BiLSTM) and Transformer based models. We occupied rank tenth in the emotion detection task by scoring a Macro F1-Score of 0.2717, validating the efficacy of our implemented approaches for small and imbalanced datasets with mixed categories of target emotions.