Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labels.
Zebin YouYong ZhongFan BaoJiacheng SunChongxuan LiJun ZhuPublished in: NeurIPS (2023)
Keyphrases
- diffusion models
- semi supervised
- label propagation
- diffusion model
- unlabeled samples
- pairwise constraints
- pairwise
- label information
- information diffusion
- semi supervised learning
- labeled data
- learning environment
- unlabeled data
- social networks
- class labels
- active learning
- training data
- supervised learning
- learning process
- influence maximization
- image data