Deadlines

There are per-lecture deadlines for paper reviews, and specific deadlines for paper presentations and research project.

For every lecture (via email to Rachit): You will need to submit a "constructive" review for each paper in the reading list of the corresponding lecture. The deadline for submitting the review is 11:59PM of the day before the lecture.

Since the objective of this class is to have constructive discussions on the assigned readings, late review submissions will not be accepted.

Slides for paper presentation (via email to Rachit) are due one week in advance. That is, for Monday lecture, you should have sent the slides to Rachit by Monday of the previous week.

[09/15] First report (via email to Rachit): If you choose to do a research project, you must submit a short (at most 2 page) description of your chosen problem and related work by 09/15. If you choose to write a survey, you must submit a list of papers that you plan to base your survey on by 09/15.

[10/20] Mid-term report (via email to Rachit): A mid-term report on your chosen research project or survey area is due by 10/20. For research projects, this report should outline the research problem, related work, and progress made so far (technique/system/algorithms) on the project. For surveys, this report should outline a list of results, a summary of individual techniques used to achieve respective results, and a few open problem.

[12/09] Final report (via email to Rachit): A final report on your chosen research project or survey is due by 12/09.

Course Grade

Your class grade will be based on the following four components:

Paper Reviews (25%): You are expected to read the papers before the lecture and write a review of the paper (see Resources section above).

Please write "constructive" reviews. Here is a rough outline.

  • Summary of problem being solved (1-2 lines)
  • Why is the problem interesting? Perhaps, its a new problem? Perhaps, its an old problem but with a new twist (e.g., new workloads, new environment, new hardware)? Perhaps, its just an old classical problem? (1-2 lines)
  • What are the main insights in the proposed solution? What is the main technical contribution? How does the solution advance the state-of-the-art? (4-5 lines)
  • Can you think of cases where the proposed solution may not work well? (4-5 lines)
  • What are the next few problems that you would solve in this space? What do you think is the holy grail in this direction? (5-10 lines)

Paper Presentations (15%): You are expected to present 5-6 papers during the semester (based on enrollment).

Please plan for 35 minutes. The paper presentations must provide a broad overview of the space (corresponding topic) and should cover all the compulsory and recommended readings. Here is a rough outline.

  • What is the problem being solved? (2-3 slides)
  • Why is the problem interesting? Perhaps, its a new problem? Perhaps, its an old problem but with a new twist (e.g., new workloads, new environment, new hardware)? Perhaps, its just an old classical problem? (1-2 slides)
  • What is the most related work and how is this paper different? (3-4 slides)
  • What is the main technical contribution? What are the main insights used to build the proposed solution? How does the solution advance the state-of-the-art? (2-3 slides)
  • Techniques used in the paper to solve the problem (4-5 slides)
  • Can you think of cases where the proposed solution may not work well? (2-3 slides)
  • What are the next few problems that you would solve in this space? What do you think is the holy grail in this direction? (4-5 slides)

Class Participation (10%): You are expected to actively participate in class discussions.

Research project or Survey (50%): Students will work on research projects (either alone, or in a team of two) or write a detailed survey (alone). For project expectations, talk to Rachit.

Academic Integrity

As with all other classes at Cornell, you are expected to maintain a high level of ethical standards and integrity in this course. This means that all work you submit must be the result of your own individual effort. You may discuss the papers with other students in the class, but you may not collaborate on the actual writing of the reviews. No part of the review may be copied from or be based on text on the web. Even higher standards apply for research projects and surveys, where plagiarism is a more well defined academic integrity violation.

Any violations of the academic integrity code will be penalized according to the Cornell Academic Integrity Policy, and may result in failure in the course, suspension, or expulsion from the university.

Here is a simple tip to avoid any problems: do NOT cheat. You know it when you are cheating! It so happens that when you are cheating, we know it too!