BPT-PLR: A Balanced Partitioning and Training Framework with Pseudo-Label Relaxed Contrastive Loss for Noisy Label Learning.
Qian ZhangGe JinYi ZhuHongjian WeiQiu ChenPublished in: Entropy (2024)
Keyphrases
- learning tasks
- multi label
- supervised learning
- learning problems
- multi task
- learning algorithm
- learning machines
- graph cuts
- multi label learning
- unsupervised learning
- training examples
- probabilistic model
- learning stage
- learning speed
- label ranking
- computer based training
- learning mechanism
- labeling effort
- object detectors
- learning scheme
- learning systems
- online learning
- object detection
- multi class
- active learning
- decision trees