Resources

Adapted from Lori Pollock's “Research Proposal Learning Experience” course assignment

The goal is to identify the world for your topic, i.e., search for all papers relevant to your topic. You won't read the papers but instead determine relevance based on title and abstract. Search in digital libraries (e.g., ACM, IEEE) and in recent conferences and workshops that cover your topic to find papers. Then, use the bibliographies of recent papers to identify earlier relevant papers. In many digital libraries, you can also see what papers the paper references and who references the given paper. There are also other helpful tools to use for spidering the bibliography, including citeseer and Google Scholar. Use Zotero or another reference management software to organize your papers into collections, e.g., earlier foundation papers on the topic, some are solving subproblem X, some solve subproblem Y, … You can also create tags for the papers.

The result should be

1. one sentence describing the overall topic/research problem you are investigating
2. a paragraph explaining how you performed your search (as precisely stated as possible)
3. a nicely formatted reference list. Consider using Bibtex and Latex, perhaps in Overleaf
4. a paragraph or enumerated list that cites the papers in your list and describes the classification of those papers in some way.

The goal is to have an organized view of the world, not just a long list of papers on the topic, which will help in writing a survey

A literature review should help you understand the timeline, overall contributions, relative merits and limitations of the work embodied in the state-of-the-art in your topic. You need to read only the abstract, introduction, related work, and conclusions sections of each paper. Do this reading in chronological order (or reverse chronological order) of paper publication dates to obtain some sense of how the research has evolved over the years.

Then, develop an outline where you have grouped the papers focusing on very similar problems, and then have a section of the outline for each paper. For each paper, be sure to include a subpart for problem addressed, contribution, findings of any evaluation of the contribution, and limitations. Just one sentence for each of these items is needed. Thus, your outline should look like:

1. Subtopic 1
   1. Paper 1: title and authors
      1. One sentence on the specific problem addressed, Contribution.
      2. One sentence on the findings of any evaluation and the limitations
   2. Paper 2: title and authors
      1. …
   3. …
2. Subtopic 2 …

This outline should be in plain text so it is easy to insert into a latex file to start writing.

First, read some Background and Related Work sections of papers to see how they are written. A good literature survey does not just write a separate paragraph on every paper written in the field in any order you want. Rather,

• A good literature survey starts with a background section that familiarizes a computer science reader to the basic topic area, such as testing web applications – what constitutes a web application, how are they characterized, examples from real life,…
• A good literature survey tries to group papers addressing similar problems and discuss and compare them together.
• A good literature survey also presents the papers typically in some chronological order within each problem identified.
• The most relevant papers to what you want to focus on are presented first, and then other papers that deal with problems related, but not so relevant, are discussed very briefly and sometimes only cited as a group with a single sentence. So, paragraphs are ordered from most relevant work to least relevant work to your chosen problem of interest.
• A good literature survey will do the following for the most relevant papers:
  • describe the overall goals/contributions of the paper
  • general approach and unique characteristics of their approach
  • restrictions/limitations of that research. What didn’t they address? Did they implement it and evaluate it?

Using your literature review outline (see above) as a template and the the suggestions above, put together your related work section. It takes time and often rewriting to put a related work section together as above.

A related work section should be 1 to 1 1/2 pages in the typical double-column conference paper format. Most are more like 1 page maximum.

Courtesy of Lori Pollock

When reading a research paper, answer the following questions:

• In one sentence in your own words, state succinctly the overall problem being addressed in this paper.
• What particular goals do these researchers have in addressing this problem?
• What contribution are they seeking to make to the state-of-the-art?

In a few sentences in your own words, what is the key insight of this group's approach to tackling the stated problem? What is their overall approach/strategy to solving the problem?

• How did the researchers evaluate their efforts?
• What conclusions did they make from their evaluation results?
• What application/useful benefit do the researchers/you see for this work?
• What limitations do the researchers mention with their approach?
• What additional limitations do you think might also occur?
• Write one interesting question to ponder with regard to this paper beyond content understanding.

• Tutorial on Research Methods in Computer Science

• ACM Digital Library
• IEEE Digital Library
• CiteSeer
• Google Scholar