CSS: Combining Self-training and Self-supervised Learning for Few-shot Dialogue State Tracking.
Haoning ZhangJunwei BaoHaipeng SunHuaishao LuoWenye LiShuguang CuiPublished in: CoRR (2022)
Keyphrases
- learning algorithm
- learning systems
- prior knowledge
- knowledge acquisition
- video sequences
- natural language
- learning process
- active learning
- natural language dialogue
- real time
- reinforcement learning
- online learning
- particle filter
- moving target
- combining multiple
- knowledge level
- dialogue system
- human computer
- dialogue management
- socio cognitive
- neural network