A Universal Part-of-Speech Tagset

By Slav Petrov, Dipanjan Das, Ryan T. McDonald. International Conference on Language Resources and Evaluation, 2011

1,071 citations recorded by the source corpus.

To facilitate future research in unsupervised induction of syntactic structure and to standardize best-practices, we propose a tagset that consists of twelve universal part-of-speech categories. In addition to the tagset, we develop a mapping from 25 different treebank tagsets to this universal set. As a result, when combined with the original treebank data, this universal tagset and mapping produce a dataset consisting of common parts-of-speech for 22 different languages. We highlight the use of this resource via three experiments, that (1) compare tagging accuracies across languages, (2) present an unsupervised grammar induction approach that does not use gold standard part-of-speech tags, and (3) use the universal tags to transfer dependency parsers between languages, achieving state-of-the-art results.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/d9674edc47fd41c27c0c6497c0e7d6bf1910d84e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Ten Pairs to Tag – Multilingual POS Tagging via Coarse Mapping between Embeddings (2016, background): “…pronoun, determiner, adpo-sition, numeral, conjunction, sentence conjunction, particle, punctuation mark, and a catch-all tag X. Note that this universal tagset contains two more tags than the traditional universal tagset proposed by Petrov et al. (2011): auxiliary verb and sentence conjunction.”
Pareto Probing: Trading-Off Accuracy and Complexity (2020, methodology): “Let us consider the POSL probing task and the case where we are using a linear model as our probabilistic probe: where t ∈ T is the target, e.g. a universal part-of-speech tag (Petrov et al., 2012), h ∈ R d is a contextual embedding and W ∈ R |T |× d is a linear projection matrix.”
Vine Pruning for Efficient Multi-Pass Dependency Parsing (2012, neutral): “In addition, we replicate each part-of-speech (POS) feature with an additional feature using coarse POS representations (Petrov et al., 2012).”
Identifying beneficial task relations for multi-task learning in deep neural networks (2017, neutral): “The number of labels.”
A Universal Framework for Inductive Transfer Parsing across Multi-typed Treebanks (2016, background): “POS tags of the heterogeneous treebanks into universal POS tags (Petrov et al., 2012).”
Towards Robust and Privacy-preserving Text Representations (2018, methodology): “Data We use the TrustPilot English POS tagged dataset (Hovy and Søgaard, 2015), which consists of 600 sentences, each labelled with both the sex and age of the author, and manually POS tagged based on the Google Universal POS tagset (Petrov et al., 2012).”
Morphological Word-Embeddings (2019, background): “For instance, in their universal POS tagset, Petrov et al. (2011) propose the coarse tag N OUN to represent all substantives.”
Syntactic Annotations for the Google Books NGram Corpus (2012, methodology): “As suggested in Petrov et al. (2012), we train on the language-speciﬁc treebank POS tags, and then map the predicted tags to the universal tags.”
Universal Dependency Annotation for Multilingual Parsing (2013, methodology): “In the context of part-of-speech tagging, universal representations, such as that of Petrov et al. (2012), have already spurred numerous examples of improved empirical cross-lingual systems (Zhang et al., 2012; Gelling et al., 2012; T¨ackstr¨om et al., 2013).”
Findings of the 2017 DiscoMT Shared Task on Cross-lingual Pronoun Prediction (2017, methodology): “We further converted the data to a lemmatized version with coarse POS tags (Petrov et al., 2012).”
What does Attention in Neural Machine Translation Pay Attention to? (2017, methodology): “We use the coarse grained universal POS tags (Petrov et al., 2012) given in Table 6.”
Knowledge Distillation for Bilingual Dictionary Induction (2017, methodology): “We use the universal POS tags, there are 12 tags in the universal POS tags (Petrov et al., 2011).”
C4Corpus: Multilingual Web-size Corpus with Free License (2016, background): “Current trends lean toward multilingual solutions, e.g., universal POS tags (Petrov et al., 2012), which requires heterogeneous corpora in multiple languages.”
A Monolingual Approach to Contextualized Word Embeddings for Mid-Resource Languages (2020, methodology): “2) and evaluate the pre-trained language models in our 5 languages, we run our experiments using the Universal Dependencies (UD)14 paradigm and its corresponding UD POS tag set (Petrov et al., 2012).”
When POS data sets don’t add up: Combatting sample bias (2014, methodology): “We use the universal tag set by Petrov et al. (2012) for this purpose.”