13th Annual Conference of the International Speech Communication Association

Portland, OR, USA
September 9-13, 2012

Coherent Topic Transition in a Conversational Agent

Daniel Macias-Galindo, Wilson Wong, Lawrence Cavedon, John Thangarajah

School of Computer Science and Information Technology, RMIT University, Melbourne, Australia

A conversational agent for entertainment and engagement requires the ability to maintain coherent conversations. We describe the use of semantic relatedness to select the next conversational fragment that an agent utters, to maximise dialogue coherence or to possibly suggest new directions for a dialogue. We compare our approach, using a specific semantic relatedness metric, to an existing nearest-context mechanism based on TF❌IDF for selecting fragments to continue a conversation. Evaluation with human judges shows that use of semantic relatedness provides improved coherence across a sample collection of generated conversations.

Index Terms: Spoken dialogue, dialogue coherence, semantic relatedness

Full Paper

Bibliographic reference.  Macias-Galindo, Daniel / Wong, Wilson / Cavedon, Lawrence / Thangarajah, John (2012): "Coherent topic transition in a conversational agent", In INTERSPEECH-2012, 743-746.