I’m Denis, a postgraduate student currently pursuing a PhD at the University of Edinburgh. I’m fortunate to conduct my research with the guidance of Rico Sennrich and Ivan Titov as part of the Statistical Machine Translation Group.

The primary goal of my PhD research is to endow neural machine translation models with useful linguistic inductive biases, through architectural design and training regime changes. By doing so, I aim to improve the overall translation quality achieved by NMT models, while also addressing specific translation challenges, such as the handling of multi-word expressions, non-literal translation, and translation of low-frequency phenomena. Beyond this, I’m also interested in low-resource neural machine translation, robustness, text generation, and common sense reasoning.

My current CV can be found here. In the future, I intend to publish my personal thoughts on language-related topics and short summaries of interesting research papers on this blog.

News

I attended WeCNLP2019! Menlo Park, USA | 06.09.2019

  • Impressions: High industry representation, excellent speakers, engaging panel discussion. Definitely worth attending.

I attended ACL2019 and WMT19!
Florence, Italy | 28.07.2019 - 2.8.2019


I published a long paper!


I’ve started a research internship at the Information Sciences Institute (ISI)!
University of Southern California, USA | 10.6.2019 - 7.9.2019


I attended the 13th Machine Translation Marathon 2018!
Charles University, Czech Republic | 3.9.2018 - 8.9.2018

  • Attended lectures and workshops on recent developments in MT
  • Implemented a hierarchical character-to-word decoder as part of the week-long hackathon

I attended the Microsoft Research AI Summer School 2018!
Microsoft Research Cambridge, UK | 2.7.2018 - 6.7.2018

  • Attended lectures and workshops at MSR Cambridge as one of 100 invited PhD students

Education

Ph.D. in Informatics
University of Edinburgh (ILCC), Scotland | 2018 - 2021

M.Sc. in Language Science and Technology
Saarland University, Germany | 2017

B.A. in German Studies University of Tübingen, Germany | 2012

Publications

Widening the Representation Bottleneck in Neural Machine Translation with Lexical Shortcuts. Emelin, Denis, Ivan Titov, and Rico Sennrich. Proceedings of the Fourth Conference on Machine Translation. 2019.

ELISA System Description for LoReHLT 2019, May, Jonathan et al. To be published.

The University of Edinburgh’s Submissions to the WMT18 News Translation Task,
Barry Haddow, Nikolay Bogoychev, Denis Emelin, Ulrich Germann, Roman Grundkiewicz, Kenneth Heafield, Antonio Valerio Miceli Barone, Rico Sennrich. In WMT 2018. Brussels, Belgium, 31 October, 2018.

Projects

Nematode
A light-weight, powerful neural machine translation toolkit built around the ‘Transformer’ model (competitive performance to T2T)

IDGAN
Master thesis project; A fully unsupervised system designed for automated and language-agnostic information density reduction/ simplification of natural language sentences; inconclusive results – currently on hold

Contact

mail | github | scholar | linkedin

Cheers!