Does the Order of Training Samples Matter? Improving Neural Data-to-Text Generation with Curriculum Learning.
Ernie ChangHui-Syuan YehVera DembergPublished in: EACL (2021)
Keyphrases
- training samples
- training data
- learning algorithm
- data sets
- supervised learning
- small number
- training set
- database
- high quality
- reinforcement learning
- feature space
- learning process
- test sample
- data distribution
- background knowledge
- information retrieval
- data points
- learning environment
- human subjects
- number of training samples