IIIDYT at IEST 2018: Implicit Emotion Classification With Deep Contextualized Word Representations

TL;DR

This paper presents a deep learning system using ELMo embeddings and BiLSTM networks for implicit emotion classification, achieving second place in the IEST 2018 shared task with a macro F1 score of 0.710.

Contribution

It introduces a novel combination of pre-trained contextualized word embeddings and BiLSTM for emotion classification, with ensembling to improve performance.

Findings

Achieved second place in IEST 2018 with 0.710 macro F1 score.

Demonstrated effectiveness of ELMo embeddings in emotion classification.

Ensembling multiple models improved overall performance.

Abstract

In this paper we describe our system designed for the WASSA 2018 Implicit Emotion Shared Task (IEST), which obtained 2$^{\text{nd}}$ place out of 26 teams with a test macro F1 score of $0.710$. The system is composed of a single pre-trained ELMo layer for encoding words, a Bidirectional Long-Short Memory Network BiLSTM for enriching word representations with context, a max-pooling operation for creating sentence representations from said word vectors, and a Dense Layer for projecting the sentence representations into label space. Our official submission was obtained by ensembling 6 of these models initialized with different random seeds. The code for replicating this paper is available at https://github.com/jabalazs/implicit_emotion.