Patrick Simianer is a principal research scientist at Lilt working on adaptive and interactive machine translation, and a former PhD student at the Lehrstuhl für Linguistische Informatik of Prof. Dr. Stefan Riezler at the Department of Computational Linguistics of Heidelberg University, Germany.
✎ p ät simianer ∙ de.
My ~/bin folder.
An assortment of mostly NLP related scripts for the shell.
Various snippets and templates for personal use.
Bare bones weather logger.
My little NLP library for ruby. Gem here. Name borrowed from George Kingsley Zipf.
Papers (and a patent) I contributed to, with accompanying materials, sorted by recency in descending order. My h-index is 8 according to Semantic Scholar.
Apparatus and Method for an Adaptive Neural Machine Translation System
A User-Study on Online Adaptation of Neural Machine Translation to Human Post-Edits
In Machine Translation, Volume 32, November 2018.
A Post-editing Interface for Immediate Adaptation in Statistical Machine Translation
In Proceedings of the 26th International Conference
on Computational Linguistics (System Demonstrations) [COLING 2016], Osaka, Japan.
Talk given at AMTA 2016 Workshop on Interacting with Machine Translation (iMT 2016).
Offline Extraction of Overlapping Phrases for Hierarchical Phrase-Based Translation
In Proceedings of the International Workshop on Spoken Language Translation (IWSLT 2014), Lake Tahoe, United States.
Online Adaptation to Post-edits for Phrase-based Statistical Machine Translation
In Machine Translation, Volume 28, December 2014.
Multi-Task Learning for Improved Discriminative Training in SMT
In Proceedings of the Eighth Workshop on Statistical Machine Translation (WMT13), Sofia, Bulgaria.
The HDU Discriminative SMT System for Constrained Data PatentMT at NTCIR10
In Proceedings of the 10th NTCIR Workshop Meeting on Evaluation of Information Access Technologies: Information Retrieval, Question Answering and Cross-Lingual Information Access, PatentMT, Tokyo, Japan.
Tuning SMT on the Training Set
Entwicklung einer webbasierten Suchmaschine und Evaluation der Bedeutung von Parametermodifikationen im Vektorraummodell
Reranker for moses that does bold updates only. Used for the 2014 MT Journal paper and 2013 MTSUMMIT paper.
Simple tuning technique implemented for cdec.
Much of the code was originally written at the machine translation marathon 2011 at Fondazione Bruno Kessler (FBK) in Trento, Italy. My development fork can be found here which
is was regularly merged into the original repo. Example data (checksum). Original slides by Stefan Riezler. Used in my Master's thesis, our ACL'12 paper, the NTCIR 10 paper, the WMT13 paper, the IWSLT 13 and 15 papers, and the user study on adaptation of statistical machine translation.
Interface for exploring learning from post-editing. Usable in conjunction with a graphical derivation editor for outputs of hierarchical phrase-based machine translation outputs. Used for two user studies for statistical- and neural machine translation.
Implementation of the one-shot- and zero-shot recall metrics proposed in the NAACL'19 paper.
Code for grounded learning for SMT. Used for the ACL 2014 paper.
I have served as reviewer for the following conferences and workshops.