Emotion recognition in the text has gotten significant attention from researchers recently. However, current models struggle with fine-grained emotion analysis, for instance, tagging emotion cause in text.
As emotion recognition is closely related to cause detection, a recent study proposes a framework for modeling them jointly. Furthermore, the study tries to combine adapted knowledge models which are trained to use common-sense knowledge with pre-trained language models.
The model yields performance gains on both emotion classification and emotion cause tagging. It is also shown that common-sense knowledge helps language models pare down the space of plausible outputs to those that are most commonly selected by human annotators. In the future, researchers hope to apply the models to other tasks, like detecting the experiencer or the target of an emotion.
Detecting what emotions are expressed in text is a well-studied problem in natural language processing. However, research on finer grained emotion analysis such as what causes an emotion is still in its infancy. We present solutions that tackle both emotion recognition and emotion cause detection in a joint fashion. Considering that common-sense knowledge plays an important role in understanding implicitly expressed emotions and the reasons for those emotions, we propose novel methods that combine common-sense knowledge via adapted knowledge models with multi-task learning to perform joint emotion classification and emotion cause tagging. We show performance improvement on both tasks when including common-sense reasoning and a multitask framework. We provide a thorough analysis to gain insights into model performance.
Research paper: Turcan, E., Wang, S., Anubhai, R., Bhattacharjee, K., Al-Onaizan, Y., and Muresan, S., “Multi-Task Learning and Adapted Knowledge Models for Emotion-Cause Extraction”, 2021. Link: https://arxiv.org/abs/2106.09790