Joint Energy-based Model Training for Better Calibrated Natural Language Understanding Models.
Tianxing HeBryan McCannCaiming XiongEhsan Hosseini-AslPublished in: EACL (2021)
Keyphrases
- probabilistic model
- natural language understanding
- statistical models
- experimental data
- language understanding
- linear model
- statistical model
- parameter estimation
- neural network model
- computational models
- high level
- parametric models
- classification models
- information retrieval
- co occurrence
- learned models
- dialogue system
- autoregressive
- prior knowledge
- probability distribution
- natural language processing
- computational model
- intelligent systems
- domain specific