Slav Petrov

4 paper cards in Mova's library · 2006–2018

Slav Petrov wrote 4 of the papers Mova reads, in 2006–2018, most of them on Universal Dependencies, Treebanks, Part-of-speech tagging; often with Daniel Zeman, Filip Ginter, Jan Hajič.

The profile is the library's match of the name; two people with one name are not always told apart.

Papers

CoNLL 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies (2018)
This overview describes the second CoNLL shared task on parsing raw text into UD. Systems start from raw text, with no gold segmentation or tags, and are scored on 82 test treebanks, with macro-averaged LAS and two new metrics, MLAS (labels plus morphological features) and BLEX (lemmas plus labels). Eight new…
CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies (2017)
The shared task asks systems to parse raw text, with no gold tokenization, sentence splitting or tags, into UD trees for 49 languages.
Universal Dependencies v1: A Multilingual Treebank Collection (2016)
The paper is the reference for the first UD release. It sets out the design principles: dependency-based and lexicalist annotation, the recoverability principle (every syntactic word must map back to the raw text), content words as the backbone of the tree with function words as dependents, and a closed inventory of…
Learning Accurate, Compact, and Interpretable Tree Annotation (2006)
Hand-made tree annotations, which split a category such as NP into subject and object NPs, raise parser accuracy, but they are made by people for one language.