Incremental Relabeling for Active Learning with Noisy Crowdsourced Annotations.
Liyue ZhaoGita SukthankarRahul SukthankarPublished in: SocialCom/PASSAT (2011)
Keyphrases
- active learning
- low quality
- batch mode
- label noise
- incremental learning
- semi supervised
- random sampling
- machine learning
- semantic annotation
- learning strategies
- metadata
- active learning strategies
- learning algorithm
- noisy environments
- incremental version
- noisy data
- image annotation
- data driven
- semi supervised learning
- labeled data
- experimental design
- supervised learning
- relevance feedback
- training set
- ground truth labels
- imbalanced data classification
- missing data
- incremental clustering
- annotation tool
- crowd sourced
- search engine
- pool based active learning