Peng Xu, Zihan Liu, Genta Indra Winata, Zhaojiang Lin, Pascale Fung
Most emotion recognition methods tackle the emotion understanding task by considering individual emotion independently while ignoring their fuzziness nature and the interconnections among them. In this paper, we explore how emotion correlations can be captured and help different classification tasks. We propose EmoGraph that captures the dependencies among different emotions through graph networks. These graphs are constructed by leveraging the co-occurrence statistics among different emotion categories. Empirical results on two multi-label classification datasets demonstrate that EmoGraph outperforms strong baselines, especially for macro-F1. An additional experiment illustrates the captured emotion correlations can also benefit a single-label classification task.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Text Classification | SemEval 2018 Task 1E-c | Accuracy | 0.589 | BERT-GCN |
| Text Classification | SemEval 2018 Task 1E-c | Macro-F1 | 0.563 | BERT-GCN |
| Text Classification | SemEval 2018 Task 1E-c | Micro-F1 | 0.707 | BERT-GCN |
| Emotion Classification | SemEval 2018 Task 1E-c | Accuracy | 0.589 | BERT-GCN |
| Emotion Classification | SemEval 2018 Task 1E-c | Macro-F1 | 0.563 | BERT-GCN |
| Emotion Classification | SemEval 2018 Task 1E-c | Micro-F1 | 0.707 | BERT-GCN |
| Classification | SemEval 2018 Task 1E-c | Accuracy | 0.589 | BERT-GCN |
| Classification | SemEval 2018 Task 1E-c | Macro-F1 | 0.563 | BERT-GCN |
| Classification | SemEval 2018 Task 1E-c | Micro-F1 | 0.707 | BERT-GCN |