Treniranje strukturnih klasifikatora za različite funkcije gubitaka sa primenom na probleme klasifikovanja sekvenci
Training Structured Classifiers for Different Loss Functions with the Application to Sequence Labeling Problems
Faculty:University of Niš, Faculty of Science and Mathematics
MetadataShow full item record
This thesis presents algorithms for training structured classifiers over different loss functions. It introduces a new primal subgradient method for the optimization of average sum loss, several extensions of max-margin classifiers to the k-best case, a sequential dual method for the structured ramp loss optimization, It considers different decoding algorithms over semirings and presents an organization of a two-structured-model committee in order to improve the results. It includes theoretical analysis of introduced algorithms, as well as experimental results on sequence labelling problems in natural language processing.
Keywords:Računarske nauke, strukturno učenje, klasifikacija sekvenci; structured learning, sequence labeling