ICS 270A: Introduction to Artificial Intelligence

Winter Quarter, 1999


Reaction Papers

To achieve success in academia, you must be able to read, understand, and critique articles from the primary scientific literature. Very little scientific progress is made without considering what others have done before. Only by reviewing the literature critically can a student begin to find areas in the field that are ripe for additional research.

Although the primary focus of ICS 270 is to cover a breadth of introductory material, the class includes some critical readings of a few seminal articles in the field. To demonstrate that you have both read and understand the assigned material, you must turn in a reaction paper for each article read. These papers may be informal, and they should be about two pages long.

A reaction paper should be more than a simple summary of the material; it should contain your opinion or reaction to reading the material. This may take on a variety of forms: you may compare the work to other related material (including the Russell & Norvig text); you may hypothesize about ways in which the work could have been improved; you may think about ways to expand on the work, or extend it to cover new domains; or you may argue against the work, questioning its assumptions, or value. I have tried to pick articles that are controversial in one way or another, which I hope will make it easier for even new students of AI to criticize or argue with these works.

Reaction papers will be graded primarily on your communication skills, and secondarily on the content and originality of your ideas. The ability to express oneself clearly is an important skill in graduate school (and in life!). Therefore, your paper, like all good essays, should include an introductory paragraph stating your main premise, a body where your detail your ideas, and a brief concluding paragraph. Although the reaction paper should not summarize the entire article, it should include enough information about the article to make your ideas or criticisms well-grounded. Assume your readers have also read the article, but that they need a quick reminder about any details you wish to discuss.

Although all writing rules are subjective and heuristic, here are a few editing pet-peeves of mine:

1) Use active voice. Passive tense sounds amorphous and wishy-washy. The use of first person ("I") is appropriate for these papers.

2) Avoid overly complex sentences. In scientific writing, these often are a death-knell for comprehension. If you do have a complex sentence, make certain it uses parallel construction.

3) Use paragraphs appropriately as partitions for your ideas. Each paragraph should generally have at least three sentences.

4) Obviously, correct spelling and complete sentences are also appreciated.

Good luck, and remember that reaction papers may not be handed in late (see grading information).


last updated Jan 2, 1999 by John Gennari