论文信息 - POS Tagging For Resource Poor Languages Through Feature Projection

POS Tagging For Resource Poor Languages Through Feature Projection

We present an approach for POS tagging with out any labeled data. Our method requires translated sentences from a pair of languages. We used feature transfer from a resource rich language to resource poor languages. Across 8 different Indian Languages, we achieved encouraging accuracies without any knowledge of the target language and any human annotation. This will help us in creating annotated corpora for resource poor Indian languages.

[1] Meher Vijay Yeleti. Improving statistical POS tagging using Linguistic feature for Hindi and Telugu , 2022 .

[2] Slav Petrov,et al. Unsupervised Part-of-Speech Tagging with Bilingual Graph-Based Projections , 2011, ACL.

[3] Dipti Misra Sharma,et al. AnnCorra : Annotating Corpora Guidelines For POS And Chunk Annotation For Indian Languages , 2008 .

[4] Slav Petrov,et al. A Universal Part-of-Speech Tagset , 2011, LREC.

[5] David Yarowsky,et al. Inducing Multilingual Text Analysis Tools via Robust Projection across Aligned Corpora , 2001, HLT.

[6] Narayan Choudhary,et al. Creating Multilingual Parallel Corpora in Indian Languages , 2011, LTC.

[7] Quoc V. Le,et al. Sequence to Sequence Learning with Neural Networks , 2014, NIPS.

[8] Thomas Hofmann,et al. Support vector machine learning for interdependent and structured output spaces , 2004, ICML.

[9] Michael Collins,et al. Discriminative Training Methods for Hidden Markov Models: Theory and Experiments with Perceptron Algorithms , 2002, EMNLP.

[10] Mark Steedman,et al. Two Decades of Unsupervised POS Induction: How Far Have We Come? , 2010, EMNLP.