Formal Models of Active Learning from Contrastive Examples

Yuxin Chen (University of Pennsylvania) · Adish Singla (MPI-SWS) · Farnam Mansouri (University of Waterloo) · Hans Simon (Max-Planck Institute) · Sandra Zilles (zilles@cs.uregina.ca)
active learnersboolean functionsclass labelsconcept classescontrastive training examplesformal studygeometric concept classesinfluence of examplesinstance differentiationlearning algorithmslearning from contrastive examplessample complexityself-directed learningtheoretical frameworktraining pairs

Machine learning can greatly benefit from providing learning algorithms with pairs of contrastive training examples