Emotionally Specific Backchanneling

Sun, 01 Jan 2023 00:00:00 +0000

Overview

Backchannels are the small verbal and nonverbal signals that communicate attention, understanding, and emotional engagement during conversation. This project examined how emotionally specific backchannels differ between human-human interaction and social human-robot interaction.

Research Focus

Rather than assuming that models trained on human conversation can be transferred directly to robots, the study compared interaction behavior across emotional contexts, including happy and sad exchanges. The goal was to identify where robot behavior needs its own interaction model instead of a direct copy of human data.

Contribution

The findings showed meaningful differences in emotionally specific backchanneling between human-human and human-robot settings. Those differences provide design evidence for interaction models that are better calibrated to a robot’s social role, helping make feedback more context-aware, intelligible, and engaging.

Conversational AI | Pourya Shahverdi

Emotionally Specific Backchanneling

Overview

Research Focus

Contribution

Project Links