Counterfactually-Augmented SNLI Training Data Does Not Yield Better Generalization Than Unaugmented Data.
William HuangHaokun LiuSamuel R. BowmanPublished in: Insights (2020)
Keyphrases
- training data
- data sets
- test data
- synthetic data
- knowledge discovery
- data collection
- databases
- noisy data
- raw data
- high dimensional data
- data processing
- data mining techniques
- original data
- training examples
- training samples
- database
- image data
- prior knowledge
- data analysis
- supervised learning
- text classification
- data warehouse
- classification accuracy
- data sources
- attribute values
- decision trees
- learning algorithm
- machine learning
- data objects
- data mining
- incomplete data
- data samples