Graph Neural Network Policies and Imitation Learning for Multi-Domain Task-Oriented Dialogues.
Thibault CordierTanguy UrvoyFabrice LefèvreLina Maria Rojas-BarahonaPublished in: SIGDIAL (2022)
Keyphrases
- multi domain
- imitation learning
- neural network
- spoken dialogue systems
- role based access control
- dialogue system
- cross domain
- domain specific
- reinforcement learning
- random walk
- humanoid robot
- heterogeneous networks
- robotic systems
- maximum margin
- optimal policy
- structured data
- multi modal
- general purpose
- decision trees
- learning algorithm