1 of 12

Paper Title

Conference, Year

Presented by:

1

2 of 12

Motivation and Context

  • What motivates this paper?
    • Some facts and figures
  • How do we translate from the real world problem to the CS problem?
    • Eg. translate to a machine learning task

2

3 of 12

Prior Art/Related Work

  • What are the prior approaches?
  • Why do they fail? What are the flaws?
    • Why is it important to address these flaws?

3

4 of 12

Approach (intuition)

  • What is the intuition behind the approach/what is the rationale behind the approach?
    • Why is the approach likely to address the flaws in related work?

4

5 of 12

Approach I

  • Give some details about the approach using a trivial example and diagram (maybe even using some trivial animations)

5

6 of 12

Approach II

  • Explain the general principle or approach which you explained in previous slide using an example

6

7 of 12

Evaluation

  • Dataset(s)
    • Give some characteristics
  • Baselines
    • A line or two about them
  • Experimental settings
    • Controlled studies?
    • In-situ settings?
    • How many participants? Etc.
    • How cross-validation was done?
    • Metrics?

7

8 of 12

Results

  • Accuracy/Error on Metric on Dataset
  • Some plots or tables

8

9 of 12

Conclusions

  • What can we conclude from the work?

9

10 of 12

Advocate [Your Name & Roll Number]

  • The paper should be accepted because of:
    • Reason 1
    • Reason 2
    • Reason 3 ….
    • …
    • ​

10

11 of 12

Devil’s Advocate [Your Name & Roll Number]

  • The paper should be rejected because of:
    • Reason 1
    • Reason 2
    • Reason 3

11

12 of 12

Some useful links for advocate/devil’s advocate

  1. http://matt-welsh.blogspot.com/2016/04/why-i-gave-your-paper-strong-accept.html
  2. http://matt-welsh.blogspot.com/2016/04/why-i-gave-your-paper-strong-reject.html

​

12