Conference paper accepted at EMNLP2015

Title: Towards the Extraction of Customer to Customer Suggestions in Reviews

Authors: Sapna Negi and Paul Buitelaar

Venue: Conference on Empirical Methods in Natural Language Processing (EMNLP), September 17–21, 2015, Lisbon, Portugal

UNLP at DPASH2015 and Digital Repository of Ireland Official Launch

The 1st Annual Conference on Digital Preservation for the Arts, Social Sciences and Humanities (DPASSH 2015) and the official launch of Digital Repository of Ireland (DRI) happened last week (25th and 26th June) at Croke Park in Dublin.

Archives are of high importance for the Natural Language Processing community because the solutions developed by researchers can be directly applied for a better extraction of information from archives. The UNLP group received two Student Bursaries to attend the event, thus Bianca Pereira and Waqas Khawaja could participate and share information about the event.

After the Welcome and Opening Remarks by Natalie Harrower, from DRI, and the president of the Royal Irish Academy, Mary Daly, the Keynote Catriona Crowe, from National Archives of Ireland, gave a talk motivating the importance of a trusted digital archive. In the past, all social and cultural information could be stored in offices or libraries which the goal was to preserve information about the past. Although such initiatives put a huge effort on the preservation of documents, a single incident as the one occurring in Public Records Office in Dublin in 1922, where the whole building was victim of an explosion and subsequent fire, can vanish all records carefully stored. This type of situation leaves an open question:”How to preserve the information about the past in a trusted way?”. Although digital repositories may store all the information in a decentralised manner, the question still remains. And this was the guide for the whole conference under the theme: “Shaping our Legacy: Safeguarding the Social and Cultural Record”

The topics approached in the first day of the conference were around trustable infrastructures and metadata, the influence of laws and ethical codes on the creation of digital archives, and the archiving of archeological as well as contemporary data already expressed in digital format (e.g. social media). The first day of conference also had the official launch of the DRI by the TD Damien English, Minister of Skills, Research and Innovation; Muiris O’Connor, from Higher Education Authority; and Sandra Collins and Natalie Harrower from DRI. The launch started with talks about the importance of DRI to safeguard records about Irish cultural and social heritage and closed with a video about the DRI and some demos.

The second day started with a Panel discussing the main theme of the conference. The panel was composed by Brid Dooley, Head of RTE Archives; John McDonough, Director of National Archives of Ireland; Helen Hockx-Yu, Head of Web Archiving at The British Library; and Owen Conlan, Assistant Professor at the School of Computer Science and Statistics in the Trinity College Dublin. The panelists and the attendance addressed the challenges in different types of digital archives and how personalisation can help on exploring such archives.

The second day continued with sessions about Policies and archiving of relevant research data, archival of text, images, and more complex objects, and preservation of art material.

The overall agenda as well as all tweet activity generated during the event can be found at the event website:

Bianca talks about STEM careers to teenagers at St Raphael’s College

This Wednesday, May 20, Bianca visited the St Raphael’s College in Loughrea to present STEM as a possible career path for Irish students. Working more than 10 years in IT and around 5 years with Research, the goal of the talk was sharing experience on what STEM is, what are the skills needed for this area, and what is the everyday life for who chooses this career.

The audience was composed by two rooms of teenagers near to go for their Leaving Certificate Program and start their career path. The main message from the talk was “Be curious, believe in yourself, and share experiences.”.

In this talk Bianca presented the differences between Science, Technology, Engineering, and Maths. Raising that the basic for Science is the curiosity about the world, focusing that people pass through many things everyday without asking themselves ‘how does this work?’. The example was the famous tale about Isaac Newton and his simple but so important question: “why do things fall down?”.

The explanation for Technology went much further than the formal definition. Technology gives human super-powers. The power to fly using an airplane, the power to talk and see a person in the other side of the world with communication technologies, or the power to detect a cancer cell using biotechnology.

Engineering was the easiest to explain. Bianca used a video of a KickStarter campaign KAZBrella to show how a simple everyday fact can be improved by applying technology to solve a real world problem. And it instantaneously called the attention of the audience for being a problem they understand and currently have.

The Mathematics explanation was more controversial. Bianca pointed that Mathematics is more than just a set of equations, it is all about critical thinking and modelling of real world. She used a joke on how someone using a programming logic could understand a simple world problem and act completely different from the usually expected. Presenting it through a challenge could show the students how the different ways of thinking may bring to different conclusions.

Bianca finished by telling her own experience on her career path and by giving pointers for who want to give IT a try without a strong commitment for this career path.

This talk was part of the SMART Futures program, a Science Foundation Ireland program to bring professionals from industry and academy to talk about STEM career to secondary students.  The slides for the presentation can be found at SlideShare.

Conference papers accepted at EAMT

Title: Identifying main obstacles for statistical machine translation of morphologically rich South Slavic languages

Authors: Maja Popovic and Mihael Arcan

Title: Poor man’s lemmatisation for automatic error classification

Authors: Maja Popovic, Mihael Arcan, Eleftherios Avramidis, Aljoscha Burchardt and Arle Lommel

Venue: The 18th Annual Conference of the European Association for Machine Translation, Antalya, Turkey

Conference Paper accepted at *SEM

Title: Non-Orthogonal Explicit Semantic Analysis

Authors: Nitish Aggarwal, Kartik Asooja, Georgeta Bordea and Paul Buitelaar

Venue: The Forth Joint Conference on Lexical and Computational Semantics (*SEM), co-located with NAACL 2015, Denver, Colorado, USA


Member of UNLP group mentioned in Irish Times Science supplement

Insight @ NUIG received a full page in Irish Times Science supplement today. And one of our members had the opportunity to contribute to the article.

The Irish Times newspaper released today, 5th November 2014, a supplement focused on the Science week in Ireland, to happen from 9-16 November. This supplement touches on different scientific topics in Ireland encouraging young people to engage in those studies.

A full page was designated for Insight Centre for Data Analytics, showing how Data Analytics is ‘the new punk’ movement and everybody can be benefited from it. In this article, Bianca Pereira, PhD student at UNLP group, could explain about her research topic and how it fits in her ambition as researcher. Ihab Salawdeh, a Research Assistant at Insight @ NUIG, could also explain about his current work in social Media and the Social Repository of Ireland.

Bianca will be presenting the cited research topic also in an Taibhdhearc Theatre next 20th November as part of the final Threesis competition.

To read the full article please visit

Mihael gives insights on SMT at the TTT conference and at the University of Ljubljana

Mihael Arcan presents his research work at the TTT conference, giving insights on word alignment, disambiguation of words to be translated and the terminology identification and injection into the SMT system. Following up the conference presentation Mihael presents his work on SMT and terminology to the students of the Faculty of Arts (University of Ljubljana). [presentation]

Join us at ISWC2014

We are very pleased to announce that our work will be presented at ISWC 2014.

Monday 20th October 2014

  • Doctoral Consortium: Entity Linking with Multiple Knowledge Bases: an Ontology Modularization Approach, Bianca Pereira
  • Full-day tutorial: Building the Multilingual Web of Data: A Hands-on tutorial, Gabriela Vulcu (

Tuesday 21st October 2014

  • Demo: EUROSENTIMENT: Linked Data Sentiment Analysis, J. Fernando Sánchez-Rada, Gabriela Vulcu, Carlos A. Iglesias and Paul Buitelaar
  • Semantic Web Challenge: Is Brad Pitt Related to Backstreet Boys? Exploring Related Entities, Nitish Aggarwal, Kartik Asooja, Paul Buitelaar and Gabriela Vulcu

Meet us at ISWC at Riva del Garda.

Threesis: 3 slides, 3 minutes, 3 judges. And UNLP got the 3rd place.

Today Andrejs Abele and Bianca Pereira participated in the first round of the Threesis challenge in National University of Ireland, Galway (NUIG). After a trial run in front of a friendly audience at Insight NUIG last week, both students were ready for the main competition. And Bianca brought the 3rd place for the UNLP group.

The Threesis challenge consists in presenting a research topic in three slides, three minutes and in front of three judges. Each speaker is judged according to their presentation skills. Today, September 30, was the first attempt for the Heating 2 at National University of Ireland, Galway. It included the areas of Informatics, Physics, Computational Sciences, Biomedical Sciences and Engineering.

At 1pm, Andrejs opened the section with his talk “Linked Data Summarisation” where he could explain in a non technical way the need of a Linked Data summarisation for developers. After one hour of presentations, Bianca brought us the third place with the talk “What if computers could read!?” where she explained the use of links for Natural Language Understanding. The second place went for Ihab Salawdeh, also from Insight NUIG, with the title “Social Repository of Ireland: Collecting & Archiving Event-Based Social Media”. In his talk, Ihab explained the social media as the history in real time. And the winner of this Heating was Killian O’Brien with the talk “The best gift comes in small packages” talking about detection of breast cancer cells.

The three winners of this Heating will participate in the NUIG final in November in Galway City Centre.

Talk for Computer Science students in Brazil

Bianca Pereira gave an invited talk about “Named Entity Recognition and Disambiguation” for Computer Science undergraduate students from Federal University of Rio de Janeiro, Brazil. The talk used Game of Thrones as a use case for Entity Recognition and linking with a brazilian wiki about the Book Series. Her talk approached the evolution of mention recognition and disambiguation methods up to the Entity Linking problem.