EE392C Class Participation Tips

The goal of EE392C is to facilitate in depth discussions of research issues and opportunities related to the architecture and system software for polymorphic/CMP processors. The success of the class and a significant portion of your final grade depend on your class participation…  

 

In practice, every student will be an instructor with the following responsibilities:

  • Read the assigned papers before each class meeting
  • Participate in the discussion in every class meeting
  • Lead the discussion during one class meeting
  • Keep notes of one class meeting
  • Review one or two final project reports

 

The following sections provide some further advice on each activity…

Reading and Summarizing Papers

 

Our in-class discussions will be of little use if we are not prepared for them. Read the required paper in advance. If the specific topic is rather new to you, you may want to also read the background papers. If you have some extra time and like this topic, you can also study the papers for further reading.

 

When reading and summarizing a paper, you are looking for the following:

  1. The issue or problem addressed by the paper
  2. The original contributions (real or claimed, you have to check)
  3. Critique: what are the major strengths and weaknesses of the papers? Look carefully at the claims and assumptions, the experimental methodology, the analysis of data, and the presentation style
  4. Future work: what are the natural extensions or improvements to this work? Or, can we apply a similar methodology to other problems of interest?

If you are the discussion leader for a topic, you will need to write down all these in a summary that does not exceed 1 page per paper.

 

I find the following techniques helpful with studying papers:

  • Read the abstract, introduction, and conclusions sections first. This should give you a good idea of the issue addressed, the basic insight of the authors, and the results achieved.
  • Read the reset paper twice. The first pass should be quick; you should try to get a rough idea of the flow of information, development of the main idea, and analysis of data. Once these seem to make sense, read the paper again and focus on the details.
  • Underline/highlight the important parts of the paper. You can take a quick look at these points right before the class…
  • Keep notes on the paper margins about comments or questions that come to your mind as you read the paper: important insights, questionable claims or techniques, relevance to other topics, ways to improve some technique etc. This will help you a lot with the class discussion.   
  • Look up references that seem to be important or missing. In some cases, you may also want to check who and how references this paper. The CiteSeer citation engine is a great source of backwards and forward citations (citation + text).

  Participating in the Class Discussion

 

EE392C is structured around round-table discussions on advanced research topics. We will all be actively teaching each other. If you prefer lecture-based courses (like most of the 200-level courses), you should probably not take this class...   

 

You are expected to come to all the class meetings and be well prepared. Study the required papers for each topic in advance. You contribution in the discussion will be in the form of comments or questions. Here are some examples of the kind of participation we are looking for:

  • Ask a question on an issue, idea, or technique that is unclear; if you cannot understand something, there is a good chance others don’t understand it either
  • Explain an issue that seems unclear at the moment or deserves further discussion
  • Propose new ideas on the topic and attempt to develop them to concrete techniques
  • Identify important issues that we are missing at the time
  • Indicate useful or damaging interactions with other fields; use your experience in other areas of computer engineering for this purpose and keep in mind the discussions from previous class meetings  
  • Provide arguments to support or oppose a proposed idea; use qualitative or quantitative arguments to support your claims

 

In general, I want you to be active in the classroom. No comment or idea is by default a bad one, so do speak up. However, be polite to your classmates and stick to the point. This is a discussion, not a competition…

Leading a Class Discussion

 

Each student is expected to lead one class discussion. The schedule will be arranged at the beginning of the quarter. If the enrollment is high, students will be paired for this task.

 

Before for class meeting:

  1. Carefully read the required papers for the class. Discussion leaders are also encouraged to study the background and further readings.
  2. Prepare a short summary for each required paper. After the class meeting, the summaries will be integrated with the discussion notes.
  3. Prepare a short presentation to initiate the discussion. Your presentation should cover the following:
    1. The issue addressed by the required papers and how does it relates to the class theme
    2. The summary of original contributions in the paper: ideas, achieved results, conclusions…
    3. Critique of the papers: major strengths and weaknesses…

This presentation should take no more than 10 minutes. You can use graphs from the papers or other visual aids. You can use a laptop, transparencies, or draw on the blackboard, whichever you are more comfortable with. 

  1. Be prepared to explain some of the detailed issues in the paper. You can safely assume that your classmates have read the papers, so there is no need to include details in the presentation. However, you should be able to clarify or explain detailed issues, if necessary, during the class discussion. You don’t need to prepare graphs or visual aids for this purpose. 

 

 

During the class meeting:

  1. Give your introductory presentation
  2. Lead the discussion. You can do this by preparing a set of interesting questions. Here are some (generic) examples:
    • What is the future work on this topic (unanswered questions)? Are there issues that we can improve on? How?
    • Are the better ways to solve the same problem?
    • Is this problem different within the context of polymorphic/CMP architectures? How can we adopt or expand the techniques in the papers to make them useful? 
    • How do the ideas on this topic interact with the issues discussed in previous class meetings?

Remember, your goal is to lead the discussion, not to dominate it. Ask your questions or use your influence when the discussion is stuck or when you believe we are missing a very important issue. Don’t expect the discussion to always go as you planned, we will hopefully generate some unexpected yet interesting ideas.

  1. In some cases, I may intervene with your role to steer the discussion towards an specific direction or make sure that we don’t miss an important point. Don’t consider this a failure on your part. Let me know if it seems I am intervening too much…

 

After the class meeting:

Meet with the student(s) that kept notes during the discussion. Merge your paper summaries and introductory presentation material with their notes in one (electronic) document and send it to me. I will post it on the class web-site. Please use on of the following formats: PDF, PS, or MS Word.

 

If necessary, I can meet with you before the class discussion and help you with the preparation, provide further pointers, go over material, etc.

Keeping Discussion Notes

 

Each student is expected to keep notes of one class discussion. The schedule will be arranged at the beginning of the quarter. If the enrollment is high, students will be paired for this task.

 

The notes must capture the essence of the discussion: significant ideas, major arguments (pros/cons), unresolved issues, conclusions, etc. You don’t have to rewrite the papers in your notes. The discussion leader will provide summaries. However, you should keep track of comments that target a specific paper.

 

After the class discussion, you should meet with the discussion leader to produce the final discussion document. It should include the summaries of the required papers (up to 1 page per paper), the introductory presentation material, and your notes. Use the following structure for the document (more or less):

  1. Short introduction to the topic
  2. Paper #1

Summary

Comments from our discussion

  1. Paper #2

Summary

Comments from our discussion

  1. Notes on other issues we discussed
  2. Conclusions & open issues

You should send me an email with the notes as an attachment within 1 week from the class meeting. Please use on of the following formats: PDF, PS, or MS Word. In some rare cases, I may ask you to revise the notes after I review them.

 

If your notes include figures or drawings (strongly encouraged): The figures should be prepared using an electronic tool (don’t just scan hand drawings) and included in the notes document. If you use Latex to prepare the document, you can prepare drawings using xfig (available on virtually all Unix/Linux systems). Xfig can export EPS files, which are easy to incorporate in a latex document. If you use MS Word, you can either use the build-in drawing capabilities or import jpg or giff images ( xfig can export your drawing in either of the two formats).

 

Click here for an example of class discussion notes from EE482A (Spring 2002).

 

If you want to use Latex for the document preparation, start with file lect00.tex and add your notes to it. You will also need file handout.cls to compile your notes. The original version of these files was developed in CS448.

Reviewing Project Papers

 

As active members of the research community, you will often be asked to review papers submitted for publication at conferences or journals. In EE392C, you will get some practice with peer reviews. Each student will review another’s group project report.

 

The reviews will be anonymous: the groups will not be told who their reviewers were.  Your goal as a reviewer is to provide an impartial review and constructive criticism. Hence, you need to make a fair assessment of the strengths and weaknesses of the paper and provide suggestions for improvements. Avoid harsh comments in the review (remember someone else is reviewing your report at the same time…). Fair and constructive are the keywords here.  

 

I strongly suggest that you read the paper “ The Task of a Referee ” by A.J. Smith (IEEE Computer, 1990).

 

Use this ASCII form for your review.