Active learning for sequence labelling with probability re-estimation

Authors:
Dittaya Wanvarie;Hiroya Takamura;Manabu Okumura
Affiliations:
Department of Computational Intelligence and Systems Science;Precision and Intelligence Laboratory, Tokyo Institute of Technology, Yokohama City, Japan;Precision and Intelligence Laboratory, Tokyo Institute of Technology, Yokohama City, Japan
Venue:
PRICAI'10 Proceedings of the 11th Pacific Rim international conference on Trends in artificial intelligence
Year:
2010

Citing 6
Cited 0

Improving Generalization with Active Learning

Machine Learning - Special issue on structured connectionist systems
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data

ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning
Introduction to the CoNLL-2000 shared task: chunking

ConLL '00 Proceedings of the 2nd workshop on Learning language in logic and the 4th conference on Computational natural language learning - Volume 7
An analysis of active learning strategies for sequence labeling tasks

EMNLP '08 Proceedings of the Conference on Empirical Methods in Natural Language Processing
Reducing labeling effort for structured prediction tasks

AAAI'05 Proceedings of the 20th national conference on Artificial intelligence - Volume 2
Semi-supervised active learning for sequence labeling

ACL '09 Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 2 - Volume 2

Quantified Score

Hi-index	0.00

Visualization

Abstract

In sequence labelling, when the label of a token in the sequence is changed, the output probability of the other tokens in the same sequence would also change. We propose a new active learning framework for sequence labelling which take the change of probability into account. At each iteration of the proposed method, every time the human annotator manually annotates a token, the output probabilities of the other tokens in the sequence are re-estimated. This proposed method is expected to reduce the amount of human annotation required for obtaining a high labelling performance. Through experiments on the NP chunking dataset provided by CoNLL, we empirically show that the proposed method works well.