Piecewise Large Margin Learning for Partially Annotated Sequences
- 1 July 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 1, 1483-1487
Abstract
Supervised and semi-supervised sequence labeling methods require large amounts of fully annotated training sequences or exact annotations of structured outputs. The problem of learning from partially annotated sequences arises in many applications, for example, Natural Language Processing and Computational Biology. In this paper, we propose Piecewise Convex Learning from Partial Labels (PW-CLPL) which is an effective discriminative structured learning method for sequence labeling as global training is intractable for partially annotated sequences. A small number of constraints is reformulated for the optimization, which improve the efficiency of parameters learning. Experimental results on the reconstructed dataset CoNLL-2000 show the effectiveness of the proposed model in the setting of partial annotations.Keywords
This publication has 8 references indexed in Scilit:
- The competitive information spreading over multiplex social networksPhysica A: Statistical Mechanics and its Applications, 2018
- Token and Type Constraints for Cross-Lingual Part-of-Speech TaggingTransactions of the Association for Computational Linguistics, 2013
- Clinical entity recognition using structural support vector machines with rich features2012
- LIBSVMACM Transactions on Intelligent Systems and Technology, 2011
- Efficient piecewise learning for conditional random fields2010
- Semi-supervised active learning for sequence labeling2009
- Semi-supervised conditional random fields for improved sequence segmentation and labeling2006
- Biomedical named entity recognition using conditional random fields and rich feature sets2004