Preserving Commonsense Knowledge from Pre-trained Language Models via Causal Inference.
Junhao ZhengQianli MaShengjie QiuYue WuPeitian MaJunlong LiuHuawen FengXichen ShangHaibin ChenPublished in: CoRR (2023)
Keyphrases
- language model
- commonsense knowledge
- causal inference
- pre trained
- training data
- information retrieval
- speech recognition
- causal models
- probabilistic model
- discrete data
- training examples
- automated reasoning
- nonmonotonic reasoning
- logic programming
- supervised learning
- semi supervised
- knowledge representation
- probability distribution
- decision trees