Not All Demonstration Examples are Equally Beneficial: Reweighting Demonstration Examples for In-Context Learning.
Zhe YangDamai DaiPeiyi WangZhifang SuiPublished in: EMNLP (Findings) (2023)
Keyphrases
- learning contexts
- learning process
- positive examples
- learning systems
- learning algorithm
- prior knowledge
- active learning
- learning models
- mobile learning
- sample selection bias
- data sets
- context dependent
- incremental learning
- learning problems
- training examples
- contextual information
- supervised learning
- semi supervised
- clustering algorithm