Description

Students are to perform in-depth study of a topic of their choice in advanced databases. The project can be an investigation of open problems or implementation of a recent state-of-the-art technique. The purpose of the assignment is twofold:

To create an opportunity so that acquired knowledge and skills are put to use in implementing advanced techniques in a collaborative environment;
To provide students upon completion of the course with a head start in conducting research in advanced databases.

The results of group projects will be made available to all students at the end of the semester.

Groups

The project will be done in groups of three. Students are free to select members among classmates to join their group. Each member is expected to be an active participant and to contribute equally to the team’s work. Any change in the group composition during the course needs approval from the instructor.

Topics

Topics investigated:

Open problems in advanced databases (complexity of the problems, suggested solutions, funding opportunities, background needed, …)
Standard datasets (time series, image databases, video databases, …) to evaluate proposed techniques, and performance metrics
Implementation of one of the recent state-of-the-art techniques, including but not limited to:

Multimedia system
Multidimensional indexing technique for high dimensional data
P2P
Multimedia processing
Multimedia delivery
Data mining
Databases
Multimedia retrieval
Bioinformatics
Data compression
Pervasive computing

Group members should get together as soon as possible and select three topics that the group feels capable of performing an in-depth investigation and producing results. A preference score and the group’s credential are given to each of the three topics. Effort will be made to assign each group with its top choice, if possible.

Sources

The technique of choice should be a state-of-the-art technique in the area of interest.
Survey papers should be from prestigious conference proceedings or journals, such as ACM Sigmod, ICDE, VLDB, ACM MM, INFOCOM.
Your original ideas, or references must be prominently cited.

Deliverables

Source code (preferably in ANSI C++) that can be compiled on UNIX and on Windows, two sets of source code if necessary.
Make files
Test input and expected output (preferably in text format)
Readme file explaining how to set up, compile, use the system interface, run the system, and interpret the results (e.g., statistics)
Known problems, if any

Final report

A typed 8-10 page report (double-spaced)
The report will include a summary which is self-contained and should provide a general idea of how the problem is addressed.
Survey of related literature
Steps in implementing the technique, modifications (e.g., to collect various statistics)
Comments about the results reported in the paper
Suggestion of how to improve the technique

Evaluation

The system will be evaluated by the GTA or by another group. The evaluation is intended to ensure, when the instructions are faithfully followed, the system can be compiled and produces correct output and handles errors well. Working systems will be made available on the class’ website at the end of the semester.

Grade formula: weighted peer evaluation and group evaluation.

Group evaluation (core grade):

Grade is based on the significance of the technique, final report and evaluation.
All members have the same core grade.

Peer evaluation:

Percentage of each member's contribution judged by the other members
Kept confidential
Decides the grade each group member receives from the core grade

Innovations encouraged

Students are always encouraged to propose new ideas. It is not uncommon that the investigation results in a new innovative proposal to address the problem of interest. Check with the instructor on how to proceed in this case. A successful proposal would result in a high grade for the entire group, presentation of the approach (instead of presentations of individual projects by members), and a paper!