Ana Cristina Mendes graduated in Informatics and Computer Science Engineering from Instituto Superior Técnico (IST), Technical University of Lisbon, Portugal. In 2006, she was introduced to the L²F group and, in particular, to Luísa Coheur, her MSc advisor. Later, in 2007, she received the MSc degree from IST, her thesis being focused on the development of a Question-Answering system for Portuguese: QA@L²F. In the same year, she became a member of the Tecnovoz project, in which she belonged to the dialogue systems development team. In 2008, she enrolled in a Ph.D. program, again under the advisory of Luísa Coheur.
INESC ID Information
- Ana Cristina Mendes, Luísa Coheur, João Pedro Carlos Gomes da Silva, Hugo Rodrigues, Just Ask A Multi-Pronged Approach to Question Answering, International Journal on Artificial Intelligence Tools, World Scientific Publishing Company, vol. 22, n. 1, pages 34, doi: 10.1142/S0218213012500364, http://www.worldscientific.com/doi/abs/10.1142/S02, February 2013
- João Pedro Carlos Gomes da Silva, Luísa Coheur, Ana Cristina Mendes, Andreas Wichert, From symbolic to sub-symbolic information in question classification, Artificial Intelligence Review, Kluwer Academic Publishers Norwell, MA, USA , vol. 35, n. 2, pages 137-154, doi: 10.1007/s10462-010-9188-4, November 2010
- Sérgio dos Santos Lopes Curto, Ana Cristina Mendes, Pedro dos Santos Lopes Curto, Luísa Coheur, Angela Costa, JUST.ASK, a QA System that Learns to Answer New Questions from Previous Interactions, In LREC 2014, Ninth International Conference on Language Resources and Evaluation, European Language Resources Association (ELRA), Iceland, May 2014
- Hugo Rodrigues, Luísa Coheur, Ana Cristina Mendes, Ricardo Ribeiro, David Martins de Matos, Testing lexical approaches in QA4MRE, In CLEF 2012, Online Working Notes, Roma, Italy, September 2012
- Pedro Fialho, Sérgio dos Santos Lopes Curto, Ana Cristina Mendes, Luísa Coheur, Extending a wordnet framework for simplicity and scalability, In Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC12), European Language Resources Association (ELRA), Istanbul, Turkey, May 2012
- Ana Cristina Mendes, Sérgio dos Santos Lopes Curto, Luísa Coheur, Bootstrapping Multiple-Choice Tests with The-Mentor, In 12th International Conference, CICLing 2011, Springer-Verlag, vol. 6608, series Lecture Notes in Computer Science, pages 451-562, Tokyo, February 2011
- Luísa Coheur, Ana Cristina Mendes, João Miguel Rodrigues da Cunha Guimarães, Nuno J. Mamede, Ricardo Ribeiro, Question Interpretation in QA@L2F, In Evaluating Systems for Multilingual and Multimodal Information Access, Springer Berlin / Heidelberg, vol. 5706, series Lecture Notes in Computer Science, pages 377-384, doi: 10.1007/978-3-642-04447-2_44, September 2009
- Ana Cristina Mendes, Luísa Coheur, Nuno J. Mamede, Ricardo Ribeiro, David Martins de Matos, Fernando Batista, QA@L2F, first steps at QA@CLEF, Springer-Verlag, vol. 5152, series Lecture Notes in Computer Science, September 2008
- Filipe Martins, Ana Cristina Mendes, Joana Paulo Pardal, Nuno J. Mamede, João Paulo da Silva Neto, Using System Expectations to manage user interactions, In The International Conference on Computational Processing of Portuguese (PROPOR 2008), Springer-Verlag, vol. 5190, series Lecture Notes in Computer Science, pages 240-243, doi: 10.1007/978-3-540-85980-2_29, Aveiro, Portugal, September 2008
- Filipe Martins, Ana Cristina Mendes, Márcio Viveiros, Joana Paulo Pardal, Pedro Arez, Nuno J. Mamede, João Paulo da Silva Neto, Reengineering a domain-independent framework for Spoken Dialogue Systems, In Software Engineering, Testing, and Quality Assurance for Natural Language Processing, an ACL 2008 Workshop, Columbus, Ohio, USA, June 2008
- Ana Cristina Mendes, Luísa Coheur, Nuno J. Mamede, QA@L2F: First Steps, In SDIA - Simpósio Doutoral em Inteligência Artificial, APPIA - Associação Portuguesa para a Inteligência Artificial, pages 13-20, Guimarães, Portugal, December 2007
- Ana Cristina Mendes, Luísa Coheur, Nuno J. Mamede, Question-Answering using a Constraint Relaxation Mechanism, In CLIN (Computational Linguistics in the Netherlands), Presentation, December 2007
- Ana Cristina Mendes, Luísa Coheur, Nuno J. Mamede, Luís Carlos da Silva Romão, João Miguel Sanches Loureiro, Ricardo Ribeiro, Fernando Batista, David Martins de Matos, QA@L2F@QA@CLEF, In Working Notes for the CLEF 2007 Workshop, September 2007
- Ana Cristina Mendes, Luísa Coheur, Sérgio dos Santos Lopes Curto, Learning to answer questions, Tech. Rep. 14 / 2012 INESC-ID Lisboa, May 2012
- Filipe Martins, Ana Cristina Mendes, Nuno J. Mamede, API da Plataforma de gestão de diálogo DIGA, Tech. Rep. 51 / 2008 INESC-ID Lisboa, September 2008
QA@L2F is the question-answering system presented in this thesis. Its architecture relies on three modules with different goals: corpus pre-processing, question analysis and interpretation and final answer
extraction. The first module is responsible for the natural language processing on the corpus and storing the information on databases. In the question analysis and interpretation module, the system recovers the question’s relevant
information (such as named entities) and sends it to the final answer extraction module.
The system last module is responsible for returning the correct answer to the input question. It can use one of four different strategies in order to retrieve the correct answer, which are: 1) linguistic pattern matching; 2) linguistic reordering; 3) named entities matching; and 4) brute force plus natural language processing. This module has a constraint relaxation mechanism, which allows the system to switch among strategies: if no answer is retrieved using the appropriate strategy, the system relaxes and tries to find an answer in a more flexible and less constrained way. QA@L2F was evaluated and tested on CLEF (an evaluation forum for information retrieval systems) and on an afterwards evaluation in the laboratory.