A Dynamic Speaker Model for Conversational Interactions
Individual differences in speakers are reflected in their language use as well as in their interests and opinions. Characterizing these differences can be useful in human-computer interaction, as well as analysis of human-human conversations. In this work, we introduce a neural model for learning a dynamically updated speaker embedding in a conversational context. Initial model training is unsupervised, using context-sensitive language generation as an objective, with the context being the conversation history. Further fine-tuning can leverage task-dependent supervised training. The learned neural representation of speakers is shown to be useful for content ranking in a socialbot and dialog act prediction in human-human conversations.
PDF Abstract