6350 Spring 2009 Project/Term-Paper

You can form a team of up-to three persons and you will get the same score. Each term is expected to 1) write a term paper or 3) a survey paper or 3) carry out a design/implementation project.

The ideal process of a term paper should have the following steps:

• P1. Define possible questions, i.e. preliminary proposal outlining areas of interest based on the examination of course material. Write a paragraph of about 300 words describing the topic of interests and link it to your project website.
• P2. Identify key sources, types of evidence. Include a list of sources. The potential source for the paper is a recent conference proceedings or a journal in the database area. Students should visit library, web (e.g. DBLP, ACM DL, IEEE DL for papers and amazon.com for book) and fellow students specializing in the interested research area, to check the availability of the sources before listing them. Provide summary of search results to illustrate the effort. List 8-12 papers closely related to the topic you choose with author(s), paper title, and publication forum in your project website.
• P3. Summary of key readings. This summary is more than a list of summaries for individual research papers using the hints in the section on analyzing a research paper. A possible approach is to find a common abstract problem addressed by all the paper and develop a simple classification system to compare the solutions proposed in individual papers. Look at the "literature survey" sections of the papers in the reading list (particularly those from ACM Computing Surveys) for examples. Write a 1 page summary of the key readings including the references. Link it to your project website.
• P4. Possible outline or overall structure for paper / project. Do include a justification comparing and contrasting your proposed paper / project with the summary of key readings. For a research paper, there should be a clear statement of the new aspects of your work. Write 2 page outline of your paper/project and link it to your project website.
• P5. Formal proposal: The proposal should list a methodology to validate the claims made in the contributions section for a research paper. Choose a methodology similar to the ones used in the papers in your reading list. For example, experiments, analytical methods, case study, detailed illustrative examples, prototyping and demonstration of new capability, etc. There should be a clear plan to list the steps within each methodology. For example an experimental methodology should include a description of the experiment design listing the candidates to be compared, metrics of performance, values of fixed parameters, value sets for variable parameters, benchmark datasets and computations, key assumptions, etc. Write a four page proposal including introduction, related work, and a more detailed description of your paper. Link it to your project website. Prepare 10 slides for a 15 minutes talk in class. (due April 7, before class)
• P6. Final report and presentation (due April 28, before class).

Possible ideas for a term paper (you are not limited to these):

(1) Identify a well defined problem, e.g. aggregation over sliding window, propose a new algorithm or processing paradigm (potentially better in some typical scenarios) to solve the problem, and validate your algorithm through experimental comparison with existing algorithms.

(2) Identify a potentially useful operator (for many applications) and the methods to process them. For example, “group nearest neighbor” operator is useful to help a group of people to find a common meeting place, taking into the individual travelling distance constraints.

(3) Find a dataset, identify a set of information/patterns/rules you want to get from the dataset, propose a slew of algorithms to get them. Address the performance issues, e.g. accuracy and computational time.

The ideal process of a survey paper should have the following steps:

• P1. Define possible questions, i.e. preliminary proposal outlining areas of interest based on the examination of course material. Write a paragraph of about 300 words describing the topic of interests and link it to your project website.
• P2. Identify key sources, types of evidence. Include a list of sources. The potential source for the paper is a recent conference proceedings or a journal in the database area. Students should visit library, web (e.g. DBLP, ACM DL, IEEE DL for papers and amazon.com for book) and fellow students specializing in the interested research area, to check the availability of the sources before listing them. List 12-15 papers closely related to the topic you choose with author(s), paper title, and publication forum in your project website.
• P3. Summary of key readings. This summary is more than a list of summaries for individual research papers using the hints in the section on analyzing a research paper. A possible approach is to find a common abstract problem addressed by all the paper and develop a simple classification system to compare the solutions proposed in individual papers. Look at the "literature survey" sections of the papers in the reading list (particularly those from ACM Computing Surveys) for examples. Write a 1 page summary of the key readings including the references. Link it to your project website.
• P4. Possible outline or overall structure for paper / project. A classification system of the surveyed work is important. Look at the "literature survey and our contribution" sections of the papers in the reading list for examples. Write 2 page outline of your paper and link it to your project website.
• P5. Formal proposal: A description of your classification scheme of the surveyed work should be included. Write a four page proposal including introduction, related work, and a more detailed description of your survey paper. Link it to your project website. Prepare 10 slides for a 15 minutes talk in class. (due April 7, before class )
• P6. Final report and presentation (due April 28, before class).

Possible topics for a survey paper (you are not limited to these):

(1) Concept drifting in stream data mining.

(2) Blocking operators in stream data processing.

(3) Stream clustering

(4) Processing dynamic nearest neighbor queries

The ideal process of a project should have the following steps:

• P1. Identify a dataset or a tool you would like to explore or expand. Define possible questions, i.e. preliminary proposal outlining areas of interest based on the examination of course material. Write a paragraph of about 300 words describing the topic of interests and link it to your project website.
• P2. Describe the dataset, tools, related papers, tutorials related to your project. Write a 1 page summarization on the dataset, tool, and related work for your project.
• P3. Describe the system architecture and the features of the system you plan to implement. Write a 2 page outline of your project and link it to your project website. Prepare 5 slides for a 15 minutes talk in class. . (due April 7, before class )
• P4. Implement the system and features. Document your implementation carefully.
• P5. Final report, presentation, and demo (due April 28, before class)..

Possible topics for a survey paper (you are not limited to these):

(1) Implement a new data type and operator in SECONDO database system (http://dna.fernuni-hagen.de/Secondo.html/)

(2) Find a large set of documents, cluster them incrementally, and compare the results with clustering the whole documents.

(3) Implement two to three algorithms found in published papers, e.g. GREEDY or FIFO on operator scheduling, simulate data input and compare results.

(4) Implement a simple publication/subscription system using triggers and stored functions in a traditional database system, e.g. MySQL.

Instructions on final report:

Follow the outlines from the research papers covered in the course. The report should be 6 pages or more and may have 5 sections:

• Problem Statement, Significance of the problem
• Related Work and Our Contributions
• Proposed Approach
• Validation of listed contributions (experimental, analytical)
• Conclusions and Future Work

For a survey paper, the report should be 6 pages or more that may include:

• Problem Statement, Significance of the problem
• Our Contributions (usually it is the categorization/classification of the research literature)
• A classification of the papers related to the problem. Use a concept hierarchy, figures, and diagrams if necessary.
• Summarize, classify, contrast, and compare the research literature according to your classification scheme
• A summary of the trend and future work of this line of research.
• Conclusion.

For implementation project, a demonstration paper of 6 pages or more is expected which may include:

• Introduction and the motivation of building such a system
• What the demo is about
• Conclusions and Future Work

In addition, you should also submit the implantation package with the documented source code, a readme file, datasets (if any), and any related documents useful for the project. A sample demo paper is available from: http://research.microsoft.com/~melnik/pub/adonet-demo_SIGMOD07.pdf

• Submit your final report using ‘project’ command in csp machines. If you are doing an implementation project, submit the implantation package through the ‘project’ command in csp machines as well. The class name is 6350s001 and the project name is ‘termProject’. (due April 28, before class )

Instructions on oral presentation:

• Oral presentation in the class using 15-25 transparencies. For demonstration project, 5-10 slides should be sufficient. Link your slides to your class project website (due April 28, before class ).
• Each project presentation should be limited to 30 minutes.
• Presentations on research papers should include motivation, major problems in the area, key results, open problems, and key sources. Focus on major problems and key results. Use summary figures (e.g. classification diagram for all approached to recovery in the Computing Survey paper in our readings) or tables to highlight key messages. Presentations on projects should follow the format of paper analysis. Candidate sections include motivation, problem definition, key issues and alternative ways of resolving those, related work and their limitations, your approach, validation, conclusions (key contributions), and future work (assumptions and potential extensions).
• Presentations of the survey paper should emphasize the classification of the research issues and methodologies used in the research literature to address the problem. The goal should be to provide a comprehensive overview of the current state-of-the-art of the field to the reader in an organized and digested way.
• Presentations on demos, you should walk the audience through the features of the system you developed. We can try to arrange a place where you can hook up your laptop for demonstration. Inform me ahead of time if you need it.