Markus Saers

profile picture

Research Associate in Computational Linguistics

Human Language Technology Center
Department of Computer Science and Engineering
The Hong Kong University of Science and Technology
HKUST, Clear Water Bay, Kowloon, Hong Kong
Lab: room 2602 (lifts 27–30) ☎ +852-2358 8831
e-mail:masaers at this domain

I earned my Doctor of Philosophy degree in Computational Linguistics from Uppsala University in 2011, and was supervised by Anna Sågvall Hein, Joakim Nivre and Dekai Wu. The main theme of the thesis was linear transductions, what they are, how they compare to other transductions (finite-state, inversion and syntax-directed).

A transduction defines a relation between two languages (in the set theoretical sense of the word language), and can thus be perceived as a set of sentence pairs. With this view it is possible to define transduction grammars and transducers to generate transductions and recognize sentence pairs as members thereof. In my thesis, linear transduction grammars (ltgs) and zipper finite-state transducers (zfsts) are introduced to fill the role of grammar and transducer for linear transductions. For a variant of ltgs, an efficient algorithm for learning from raw parallel corpora is also included in the thesis. Formally, the class of linear transductions is strictly larger than finite-state transductions and strictly smaller than inversion transductions.

Research Interests

My research interests are mainly focused on machine translation. Specifically how formal grammars can be used to characterize the relation between natural languages, and how that relation can be induced from raw data.


Beloucif, Meriem, Markus Saers & Dekai Wu (2015) Improving Semantic SMT via Soft Semantic Role Label Constraints on ITG Alignments in Proceedings of Machine Translation Summit XV (MT Summit 2015), pages 333–345, Miami, Florida, October 2015.
Wu, Dekai, Chi-kiu Lo, Meriem Beloucif & Markus Saers (2014) Better Semantic Frame Based MT Evaluation via Inversion Transduction Grammars in Proceedings of the Eigth Workshop on Syntax, Semantics and Structure in Statistical Translation (SSST-8), pages 22–33, Doha, Qatar, October 2014.
Saers, Markus & Dekai Wu (2014) Ternary Segmentation for Improving Search in Top-down Induction of Segmental ITGs in Proceedings of the Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation (SSST-8), pages 48–57, Doha, Qatar, October 2014.
Wu, Dekai, Chi-kiu Lo & Markus Saers (2014) Lexical Access Preference and Constraint Strategies for Improving Multiword Expression Association within Semantic MT Evaluation in Proceedings of the 4th Workshop on Cognitive Aspects of the Lexicon (CogALex), pages 144–153, Dublin, Ireland, August 2014.
Lo, Chi-kiu, Meriem Beloucif, Markus Saers & Dekai Wu (2014) XMEANT: Better semantic MT evaluation without reference translations in Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (ACL 2014), pages 765–771, Baltimore, Maryland, June 2014.