CS 8803 DMM: Data Management and Machine Learning (Fall 2019)

Paper Review

For each class, the student must write review for every paper, and submit the paper review the night before class.

The presenter of a paper will lead the discussions of the questions and comments in the reviews he/she receives. All students are required to actively participate in the discussion.

Submission Guideline

• The review must be emailed to the Instructor and the TA by 6:00pm EST the night before the class.
  
• The title of the email must be "CS 8803 DMM Paper Review {Your GT User ID} {Title of the Paper}"
• The review content should be in the body of the email, NOT as an attachment.
  
• Late submissions will not be accepted.
  
• If you are presenting a paper for a class, then you don't need to submit any review for that class.
  
• Given unpredictable workloads students may have during the semester, you are allowed to miss at most three reviews during the semester. 

Review Structure

Each review must contain the following information:

• Summary: (An overview of the main idea and contributions in one paragraph)
• What is this paper about?
  
• What is the main contribution?
  
• Describe the main approach & results. Just facts, no opinions yet.
  
• Summarize the insights resulting from the empirical study.
• Strength: (State at least three strengths of the paper)

• Is there a new theoretical insight? Or a significant empirical advance?
• Did they solve a standing open problem? Or a good formulation for a new problem? Or a faster/better solution for an existing problem?
• Any good practical outcome (code, algorithm, etc)?
• Are the experiments well executed?
  
• Useful for the community in general?
• Other types of strengths you may think of ... 
• Weakness: (State at least three weaknesses of the paper.)
• What can be done better?
• Any missing baselines? Missing datasets?
• Any odd design choices in the algorithm not explained well?
• Quality of writing?
• Is there sufficient novelty in what they propose?
• Minor variation of previous work?
• Why should anyone care? Is the problem interesting and significant?
• Other types of strengths you may think of ...
• Two discussion points: (To be brought by you in classroom discussion)
• How does this relate to other papers we have read?
  
• What are the next research directions in this line of work?
  
• What (directly or indirectly related) new ideas did this paper give you?
  
• What would you be curious to try?
  
• What experiments can deepen our understanding?
  
• How can your own research benefit from the insights provided in the paper?
  
• Why is the paper not interesting?
  
• What is the paper’s potential impact on the field?

Grading

Each review will be evaluated on a 100 points scale:

• 20 points for summary
• 20 points for strengths
  
• 20 points for weakness
• 20 points for discussion points
• 20 points for clarity and cogent statements
• Bonus: 10 points for exceptional reviews with new insights that can inspire future work