DSc candidate (Computer Science and Computational Mathematics) - GBdI/ICMC/USP and Le2i/uB
Contacts
Grupo de Base de Dados e Imagens (GBdI)
Instituto de Ciências Matemáticas e de Computação (ICMC)
Universidade de São Paulo (USP)
Av. Trabalhador São-Carlense, 400
13566-590 - São Carlos-SP, Brasil
Phone: +55-16-3373-9677
Fax: +55-16-3371-2238
e-mail: monika at icmc dot usp dot br
Laboratoire Electronique, Informatique et Image (Le2i)
Université de Bourgogne (uB)
Aile de l'Ingénieur
21078 - Dijon, France
Phone: +333 80 39 36 55
Fax: +333 80 39 68 69
e-mail: Monica.Ferreira at u-bourgogne dot fr
Research Interests
- Query optimization
- Similarity algebra
- Similarity query
- Query by content in image databases
- Multimedia databases
Research Projects
- Doctorate
Title: Optimizing similarity queries on metric spaces meeting users' expectations.
Advisors: Prof. Dr. Caetano Traina Júnior (ICMC/USP) and Prof. Dr. Richard Chbeir (uB)
Sponsor(es): FAPESP (Process N. 2008/00210-7) and CAPES/PDEE (Process N. 2451/09-3)
Abstract: The complexity of data stored in large databases has increased at very fast paces. Hence, operations more elaborated than traditional queries are essential in order to extract all required information from the database. Therefore, the interest of the database community in similarity search has increased significantly. Two of the well-known types of similarity search are the range queries (Rq) and the k-Nearest Neighbor queries (kNNq), which, as any of the traditional query, can be sped up by indexing structures of the Database Management System (DBMS). Another way of speeding up queries is to perform query optimization. In this process, metrics about data are collected and employed to adjust the parameters of the search algorithms in each query execution. However, although the integration of similarity search into DBMS has begun to be deeply studied more recently, the query optimization has been yet developed and employed to answer just traditional queries. The execution of similarity queries, even using efficient indexing structures, tend to present higher computational cost than the execution of traditional queries. Two strategies can be applied to speed up the execution of any query, and thus are worth to employ to answer also the similarity queries. The first strategy is the query rewriting based on cost functions. The second technique is employing query external factors, such as the semantic expected by the user, to prune the answer space. This Project aims at developing techniques to integrate similarity query algorithms in the DBMS core, in order to make possible using a query optimization structure in efficient ways. Also, this project intends to gather external conditions, which represent the user's interest as a tool to offer additional information to improve query optimization.
- Master
Title: Extending SQL to support unary similary queries (Suporte a consultas por similaridade unárias em SQL).
Advisor: Prof. Dr. Caetano Traina Júnior
Sponsor: FAPESP (Process 2005/03341-7)
Abstract: Conventional operators for data comparison based on exact matching and total order relations are not appropriate to manage complex data, such as multimedia data (e.g. images, audio and large texts), time series and genetic sequences. In fact, the most important aspect to compare complex data is usually the similarity degree between instances, leading to the use of similarity operators to perform search and retrieval operations. Similarity operators can be classified as unary or as binary, respectively used to implement selection operations and joins. However, the Relation Algebra, employed in Relational Database Management Systems (DBMS), does not provide resources to express similarity search criteria. In order to fulfill this lack of support, an extension to the Relational Algebra is under development at GBdI-ICMC-USP (Grupo de Bases de Dados e Imagens), aiming to represent similarity queries in algebraic expressions. This work contributes to such an effort by dealing with unary similarity operators in Relational Algebra and by developing a similarity query optimizer for SIREN (Similarity Retrieval Engine), therefore allowing similarity queries to be answered by Relational DBMS.
Formal Education
- 02/2001 - 12/2005
BSc (Informatics), ICMC-São Carlos/SP, Universidade de São Paulo, Brazil.
- 03/2006 - 02/2008
MSc (Computer Science and Computational Mathematics), ICMC-São Carlos/SP, Universidade de São Paulo, Brazil. "Extending SQL to support unary similary queries". Advisor: Prof. Dr. Caetano Traina Júnior
- 03/2008 - today
DSc (Computer Science and Computational Mathematics), ICMC-São Carlos/SP, Universidade de São Paulo, Brazil with co-supervision of Université de Bourgogne, France. Advisors: Prof. Dr. Caetano Traina Júnior and Prof. Dr. Richard Chbeir.