• :Home
• :publications
• :teaching
• :software
• links

Data and Resources

home | cv | projects | resources

:resources for morphosyntactic parsing

• Joint Decoding
  The lattice-based decoder in [2] has been deployed in the Stanford Parser [4] and the Berkely parser [3]. Our 2008 implementation is deprecated. For joint decoding I am currently using blatt.
• Joint Training
  Trained models for RR parsing [5,6] of English, Swedish and Hebrew (alongside baseline models trained for constituency and dependency parsing) can be found here. These data were used in [7,8].
• Joint Evaluation
  The TedEval software for joint morphosyntactic evaluation [7]. The software can also be used for linguistic tree conversions, and was used to generate the training materials in [8].
• Joint Annotation
  Description of the unified morpho-syntactic dependency scheme (extending Stanford Dependencies [1]) with an application to the Hebrew data sets. [8].

:references

[1] de Marneff at al CoLing 2008 [pdf]
[2] Goldberg and Tsarfaty ACL 2008 [ pdf]
[3] Goldberg PhD Thesis 2011 [pdf]
[4] Green and Manning LREC 2010 [pdf]
[5] Tsarfaty and Sima'an CoLing 2008 [pdf]
[6] Tsarfaty PhD Thesis 2010 [pdf]
[7] Tsarfaty, Nivre, Anderson ACL 2012 [pdf]
[8] Tsarfaty ACL 2013 [pdf]

:distribution

This page does not provide IP info for the linked resources. If you have downloaded or used any of these data or resources for academic purposes, please cite the relevant work. If you use it for commercial applications, please follow the license restrictions stated therein.

Visit the publications page for additional bibliographic information. If you found any if these resources useful, it would be great if you drop me a note. I am happy to follow up on any further questions or comments you may have.