Non-personal page of Aline Marins Paes Carvalho

Address:

Aline Marins Paes
PhD Student
Universidade Federal do Rio de Janeiro
Programa de Engenharia de Sistemas
Centro de Tecnologia, Bloco H, sala 317-12, Caixa Postal 68511
Ilha do Fundao, 21945-970 - Rio de Janeiro, RJ - Brasil

Phone: +55 21 2562-8679
E-mail:
ampaes at cos dot ufrj dot br

Welcome!

I am a PhD candidate at the Department of Systems Engineering and Computer Science (PESC), filiated to the Alberto Luiz Coimbra Graduate School and Research in Engineering (COPPE) at the Federal University of Rio de Janeiro (UFRJ). The supervisors of my PhD thesis are Professor Gerson Zaverucha and Professor Vitor Santos Costa. Before starting the PhD, I received a M.Sc. in Systems Engineering and Computer Science (2005) from COPPE/UFRJ and, before that, a B.Sc. degree in Computer Science (2002) from University of the State of Rio de Janeiro (UERJ). Big Ben

From July, 2008 to July, 2009 I had been visiting the Computational Bioinformatics Laboratory of the Department of Computing at Imperial College London, under the supervision of Professor Stephen Muggleton.

Research Interests

My main research area is (Probabilistic) Logical and Relational Learning, which mainly combines ideas of Machine Learning and Knowledge Representation. Probabilistic Logic Learning (aka Probabilistic ILP and quite related to Statistical Relational Learning) is an exciting, emerging research area in the intersection of ILP and Probabilistic Learning. My PhD thesis is about the use of theory revision techniques in Probabilistic Logic Learning and ILP. I intend to apply such thecniques to solve problems in areas such as bioinformatics, games and natural language processing.

Research Topics

Development of efficient search strategies to learn and revise structure and parameters of probalistic logic models;
Development of novel search strategies which allow efficiently to induce and revise whole theories (evaluating theories globally instead of evaluating individual clauses locally);
Stochastic Local Search;
Incremental theory revision;
Predicate and hidden variable invention;
Machine Learning and Games;
Applications of PILP/SRL in Social Networks, Natural Language Learning, Biological networks and SAR domains.

Softwares

PFORTE, developed together with Kate Revoredo.
Revision of Probabilistic Logic Models from Examples, inspired on FORTE. Currently, it is able to revise Bayesian Logic Programs, also making use of novel predicate invention operators.
ASSERT (implemented in FORTE 3.0) (A Stochastic Search Engine for Revising Theories)
Revision of First-order Logic Programs from Examples using Stochastic Local Search Algorithms.
DAHVI (Discriminative Approach for Hidden Variable Introduction), developed together with Kate Revoredo
Introduction of hidden variables in Bayesian Networks from a theory revision perspective.
FORTE_MBC (implemented in FORTE 4.0), developed together with Ana Luisa Duboc, Eric Couto, Rafael Pereira and Guilherme Niedu
Revision of First-order Logic Programs from Examples, using Modes Declarations and the Bottom Clause.

Academic Service

Service in Conferences:

Reviewer for ILP 2009, 2008, 2007, 2006.
Reviewer for Discovery Science 2009.
Reviewer for ICML 2008, 2007.
Reviewer for ECML 2009, 2008, 2007, 2005.
Reviewer for KDD 2009, 2007.
Reviewer for ICDM 2007, 2006.
Reviewer for ENIA 2009, 2007, 2005.
Participation in IJCAI 2009.
Participation in ILP 2009, 2008, 2007, 2006.
Participation in ICML 2007.
Participation in SBIA 2006.
Participation in ENIA/SBC 2007, 2005.

Current Students (co-advisor, together with Gerson Zaverucha):

Eric Couto Luz e Silva (UFRJ,2008), Undergraduate Research Project.

Alumni (co-advisor, together with Gerson Zaverucha):

Ana Luisa Duboc (UFRJ, 2008), Msc in Systems Engineering and Computer Science.
Rafael Jose Goncalves Pereira (UFRJ, 2008), Undergraduate Research Project.
Guilherme Silva Niedu (UFRJ, 2008), Undergraduate Research Project.
Thiago Braga (UFRJ, 2008), Undergraduate Research Project.

Aline Paes