Data point selection for self-training

Problems for parsing morphologically rich languages are, amongst others, caused by the higher variability in structure due to less rigid word order constraints and by the higher number of different lexical forms. Both properties can result in sparse data problems for statistical parsing. We present a simple approach for addressing these issues. Our approach makes use of self-training on instances selected with regard to their similarity to the annotated data. Our similarity measure is based on the perplexity of part-of-speech trigrams of new instances measured against the annotated training data. Preliminary results show that our method outperforms a self-training setting where instances are simply selected by order of occurrence in the corpus and argue that self-training is a cheap and effective method for improving parsing accuracy for morphologically rich languages.

[1]  Amit Dubey,et al.  Statistical parsing for German: modeling syntactic properties and annotation differences , 2005 .

[2]  Anders Søgaard.  Data point selection for cross-language adaptation of dependency parsers , 2011, ACL.

[3]  Sandra Kübler.  How Do Treebank Annotation Schemes Influence Parsing Results? Or How Not to Compare Apples And Oranges , 2005 .

[4]  Eugene Charniak,et al.  Coarse-to-Fine n-Best Parsing and MaxEnt Discriminative Reranking , 2005, ACL.

[5]  Dan Klein,et al.  Learning Accurate, Compact, and Interpretable Tree Annotation , 2006, ACL.

[6]  Philip Resnik,et al.  Cross-Language Parser Adaptation between Related Languages , 2008, IJCNLP.

[7]  Eugene Charniak,et al.  Reranking and Self-Training for Parser Adaptation , 2006, ACL.

[8]  Josef van Genabith,et al.  Why is it so difficult to compare treebanks? TIGER and TüBa-D/Z revisited , 2007 .

[9]  Mary P. Harper,et al.  Self-Training PCFG Grammars with Latent Annotations Across Languages , 2009, EMNLP.

[10]  Eugene Charniak,et al.  Effective Self-Training for Parsing , 2006, NAACL.

[11]  Sandra Kübler.  The PaGe 2008 Shared Task on Parsing German , 2008 .

[12]  Ari Rappoport,et al.  Self-Training for Enhancement and Domain Adaptation of Statistical Parsers Trained on Small Datasets , 2007, ACL.

[13]  Kathrin Beck,et al.  Stylebook for the Tubingen Treebank of Written German (TuBa-D/Z) , 2012 .

[14]  Mark Steedman,et al.  Bootstrapping statistical parsers from small datasets , 2003, EACL.

[15]  Josef van Genabith,et al.  Adapting WSJ-Trained Parsers to the British National Corpus using In-Domain Self-Training , 2007, IWPT.

[16]  Helmut Schmid,et al.  Estimation of Conditional Probabilities With Decision Trees and an Application to Fine-Grained POS Tagging , 2008, COLING.

[17]  Dan Klein,et al.  Accurate Unlexicalized Parsing , 2003, ACL.

[18]  Eugene Charniak,et al.  Automatic Domain Adaptation for Parsing , 2010, NAACL.

[19]  Eugene Charniak,et al.  Statistical Parsing with a Context-Free Grammar and Word Statistics , 1997, AAAI/IAAI.

[20]  Sabine Brants,et al.  The TIGER Treebank , 2001 .

[21]  Brian Roark,et al.  MAP adaptation of stochastic grammars , 2006, Comput. Speech Lang..

[22]  Josef van Genabith,et al.  Hard Constraints for Grammatical Function Labelling , 2010, ACL.

[23]  Eugene Charniak,et al.  When is Self-Training Effective for Parsing? , 2008, COLING.