CAiRE_HKUST at SemEval-2019 Task 3: Hierarchical Attention for Dialogue Emotion Classification
This work addresses the challenge of detecting emotion in dialogue, which is an incremental improvement for natural language processing applications.
The paper tackled emotion classification in dialogue by proposing a hierarchical approach that models current emotional state based on previous latent emotions, achieving a 76.77% F1-score on the test set.
Detecting emotion from dialogue is a challenge that has not yet been extensively surveyed. One could consider the emotion of each dialogue turn to be independent, but in this paper, we introduce a hierarchical approach to classify emotion, hypothesizing that the current emotional state depends on previous latent emotions. We benchmark several feature-based classifiers using pre-trained word and emotion embeddings, state-of-the-art end-to-end neural network models, and Gaussian processes for automatic hyper-parameter search. In our experiments, hierarchical architectures consistently give significant improvements, and our best model achieves a 76.77% F1-score on the test set.