|
VLIR-HUT
Research Project |
|
NLI4DB: A Natural Language Interface for Querying Database and Automatically Generating Reports
Duration: 2 years (April 1, 2006 - March 31, 2008)
Promoter-spokesperson: | Dr. Le Thanh Huong |
Co-promoters: | Dr. Nguyen Kim Anh, Dr. Vu Tuyet Trinh |
This project aims at developing methods for querying databases using natural language (e.g., "Who is the leader of the class BK20 in the academic year 2004-2005?") and generating summary reports in different formats: tables, text, graphics, or a combination of them. The output of this research is a tool that can integrate with existing Database Management Systems (DBMSs) so as to provide a Natural Language Interface for Querying Database and Automatically Generating Reports (NLI4DB). The result is to improve quality of data processing and to reduce workload for official staffs. An overview of the architecture of the proposed tool is shown in Figure 1.
To achieve these objectives, three major technical problems are needed to solve:
P1: Design an interface where users can compose questions in natural language, which is Vietnamese in our case.
P2: Translate questions from a natural language to a query language.
P3: Generate responses to the user's questions. These may range from short answers to summary reports and may be composed in the format of graphics, text, or a combination of them, as appropriate to the query result.
This is a two year project but conceived as the initiate point of a long term program of research. The cooperation with national and international research centers will be continued to optimize research theories and results. The research results will be published in various academic fora such as journal papers, conference's papers and technical reports. This project builds on an existing (informal) international collaboration and will strengthen the cooperation between national scientists with the leading scientists in the area and in the world.
The final (long-term) goal of this project is to establish technologies and to produce a robust framework for the proposed NLI4DB, which can be used in different application areas (e.g., health services, insurance, crime investigation, etc.). Due to the limitation of time, the implementation of this project will focus on a student management database.