EmoBERTa: Speaker-Aware Emotion Recognition in Conversation with RoBERTa
This addresses the problem of accurately detecting emotions in multi-speaker dialogues for applications like chatbots and social media analysis, but it is incremental as it builds on existing RoBERTa models with simple modifications.
The paper tackled emotion recognition in conversation by proposing EmoBERTa, a method that prepends speaker names and uses separation tokens to leverage speaker-aware context, achieving new state-of-the-art results on two popular datasets.
We present EmoBERTa: Speaker-Aware Emotion Recognition in Conversation with RoBERTa, a simple yet expressive scheme of solving the ERC (emotion recognition in conversation) task. By simply prepending speaker names to utterances and inserting separation tokens between the utterances in a dialogue, EmoBERTa can learn intra- and inter- speaker states and context to predict the emotion of a current speaker, in an end-to-end manner. Our experiments show that we reach a new state of the art on the two popular ERC datasets using a basic and straight-forward approach. We've open sourced our code and models at https://github.com/tae898/erc.