Hi. I'm Daniel Duma. I am a PhD student at the University of Edinburgh, in the Institute for Language, Cognition and Computation, School of Informatics, under the supervision of Ewan Klein, Charles Sutton and Maria Liakata. I am currently based at the Alan Turing Institute.

My current PhD research deals with recommending relevant citations to the author of an academic paper. Like others before, I am framing this task as an Information Retrieval scenario but I am trying to incorporate more sophisticated Natural Language Processing techniques. One particularly important avenue of research so far has been applying functional argumentation schemes to context-based citation recommendation. Two of these schemes I am working with are CoreSC and Argumentative Zoning. My upcoming focus is query generation and keyword extraction using neural sequential models.

I am interested in all areas of NLP and AI, and I have previous experience in Natural Language Generation for Linked Data for my MSc dissertation.

Publications

  • Daniel Duma, Maria Liakata, James Ravenscroft, Amanda Clare and Ewan Klein. 2016. Rhetorical Classification of Anchor Text for Citation Recommendation. In D-Lib Magazine. [link] [pdf] [slides]
  • Daniel Duma, Charles Sutton and Ewan Klein. 2016. Context Matters: Towards Extracting a Citation’s Context Using Linguistic Features. In Proceedings of the 16th ACM/IEEE-CS Joint Conference on Digital Libraries (JCDL 2016). [paper] [poster]
  • Daniel Duma, Maria Liakata, James Ravenscroft, Amanda Clare and Ewan Klein. 2016. Applying CoreSC to context-based citation recommendation. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC-2016), 23-28 May 2016, Portorož (Slovenia) . [paper] [poster]
  • Daniel Duma and Ewan Klein. 2014. Citation Resolution: A method for evaluating context-based citation recommendation systems In Proceedings of the Association for Computational Linguistics (ACL'14). [paper]. [slides]
  • Daniel Duma and Ewan Klein. 2013. Generating Natural Language from Linked Data: Unsupervised template extraction . In Proceedings of the International Conference of Computational Semantics IWCS'13. [paper]
  • Daniel Duma. 2012. Natural Language Generation for the Semantic Web: Unsupervised Template Extraction . MSc Dissertation (distinction), University of Edinburgh. [dissertation]

Startups and projects

Scholarly With Scholarly (Scholarly Software Ltd, formerly www.scholarlyessay.com) we were aiming to create automated copy-editing software for academic writing: software that would tell you what to cut and how to reword it to fit the word limit, who to cite based on what you are saying, how to better structure your paragraphs, etc. We received investment and went through the 3-month Dotforge accelerator programme, and failed to raise sufficient funding with some 300 users. I wore many hats at Scholarly, including CEO and main product/UI designer, but I also wrote many lines of Python and Javascript.

beacon Beacon (www.wearbeacon.com) is all about personalization. We set out to make a wearable, flexible, breathable screen that you can put on any fabric. We started with a t-shirt, of which we made several prototypes and we reached the semi-finals of the Intel Make it Wearable competition and received some seed funding from the University of Edinburgh. Eventually we dropped the effort just ahead of our kickstarter when the whole team decided we had PhDs we should better get on with. I was also CEO, product designer and lead programmer on this effort, mostly for the device firmware (Wiring/C/C++) but also for the Android app (Java).

CV / Résumé

Education

2013-present PhD in Computational Linguistics & Information Retrieval (University of Edinburgh)
2011-2012 MSc Speech and Language Processing (Distinction) (University of Edinburgh)
2005-2010 BA Linguistics & English Studies (Universidad Autónoma de Madrid)
July 2009 Cambridge Certificate in English Language Teaching to Adults (CELTA) - Pass B (Westminster Kingsway College, London)
2008-2009 Full year visiting student, Goldsmiths College, University of London: 10 months (Erasmus scholarship)

Work

May-Sept 2015 Google – Software Engineering Intern. Zurich, Switzerland
2013, 2014 University of Edinburgh – Teaching Assistant, Marker, Demonstrator
Apr-Dec 2013 Scholarly Software Ltd. – Co-founder and Director.
2010 – 2011 Eton College, Windsor, UK – Spanish Assistant Teacher
2009 - 2010 Erasmus Student Network UAM – Committee member in charge of Internal Relations
2006 – 2007 Spanish Army Central Headquarters – Delphi Programmer

Languages


I speak Spanish (native), English (native) and French (if pressed).

Contact

Email me at [myname][mysurname] (no dot, no spaces) @ gmail dot com , or find me on twitter, github or linkedin.