Active Learning

[azhe825]

Re-active Learning: Active Learning with Relabeling

[azhe825]

Re-active Learning: Active Learning with Relabeling

Labels are not always perfect. (crowd sourcing...)

Whether to label a new example or to relabel an existing example.

impact sampling: label the sample that can change the classifier most.

uncertainty sampling: label the sample that the classifier is most uncertain about.

[azhe825]

Basic active learning:

Active Learning, Synthesis Lectures on Artificial Intelligence and Machine Learning (BOOK)

Software product line related: (?)

Active Learning with Clustering this one works with really huge data set and similar to what vivek has done: find representatives, use them to build hierarchical clusters (instead of decision tree).

Transfer active learning: (defect prediction?)

Active Learning with Cross-Class Knowledge Transfer

and its references.

Clustering-based active learning: (StackOverflow Data?)

Hierarchical sampling for active learning

Crowd sourcing related:

Re-active Learning: Active Learning with Relabeling

[azhe825]

Active learning for class imbalance problem 2007

Learning on the border: active learning in imbalanced data classification 2007 (Extended version of above)

Why label when you can search?: alternatives to active learning for applying human resources to build classification models under extreme class imbalance 2010 (Guided Learning)

Boosted Disagreement with QBC:

Reducing class imbalance during active learning for named entity annotation.

[azhe825]

2016:

Search Improves Label for Active Learning Beygelzimer, Alina, Daniel Hsu, John Langford, and Chicheng Zhang. "Search Improves Label for Active Learning." arXiv preprint arXiv:1602.07265 (2016).

Near-optimal Bayesian Active Learning with Correlated and Noisy Tests Chen, Yuxin, S. Hamed Hassani, and Andreas Krause. "Near-optimal Bayesian Active Learning with Correlated and Noisy Tests." arXiv preprint arXiv:1605.07334 (2016).(NIPS 2016) Noisy labels

Exponentiated Gradient Exploration for Active Learning Bouneffouf, Djallel. "Exponentiated Gradient Exploration for Active Learning." Computers 5, no. 1 (2016): 1. Explore vs. exploit

[azhe825]

SE:

Sample-based software defect prediction with active and semi-supervised learning 2012 Li, Ming, Hongyu Zhang, Rongxin Wu, and Zhi-Hua Zhou. "Sample-based software defect prediction with active and semi-supervised learning." Automated Software Engineering 19, no. 2 (2012): 201-230. CoForest, ACoForest

[Label propagation based semi-supervised learning for software defect prediction]() 2016 Zhang, Zhi-Wu, Xiao-Yuan Jing, and Tie-Jian Wang. "Label propagation based semi-supervised learning for software defect prediction." Automated Software Engineering (2016): 1-23.