Kim, Hanseob Ali, Ghazanfar Hwang, Jae-In Kim, Gerard J. Kim, Seungwon 2024-01-19T09:08:33Z 2024-01-19T09:08:33Z 2022-02-25 2021-03 https://pubs.kist.re.kr/handle/201004/113576 Virtual humans are most natural and effective when it can act out and animate verbal/gestural actions. One popular method to realize this is to infer the actions from predefined phrases. This research aims to provide a more flexible method to activate various behaviors straight from natural conversations. Our approach uses BERT as the backbone for natural language understanding and, on top of it, a jointly learned sentence classifier (SC) and entity classifier (EC). The SC classifies the input into conversation or action, and EC extracts the entities for the action. The pilot study has shown promising results with high perceived naturalness and positive experiences. English IEEE Auto-generating Virtual Human Behavior by Understanding User Contexts Conference 10.1109/VRW52623.2021.00178 1 28th IEEE Conference on Virtual Reality and 3D User Interfaces (IEEE VR), pp.591 - 592 28th IEEE Conference on Virtual Reality and 3D User Interfaces (IEEE VR) 591 592 US ELECTR NETWORK 2021-03-27 2021 IEEE CONFERENCE ON VIRTUAL REALITY AND 3D USER INTERFACES ABSTRACTS AND WORKSHOPS (VRW 2021) 000676157600171 2-s2.0-85105966792