Home

Curious Snake is an active learning framework in Python. Active learning generally refers to strategies for training classifiers interactively — i.e., allowing the classification model to request labels for examples it is ‘interested’ in, or that will best facilitate training. Those unfamiliar with active learning should consult Settle’s literature survey: http://pages.cs.wisc.edu/~bsettles/pub/settles.activelearning.pdf. This library focuses on pool-based active learning, in which there is a large pool of unlabeled examples, or instances, that can be labeled at a cost (usually time).